Wikisource:Scriptorium

From Wikisource
Jump to: navigation, search
Scriptorium
The Scriptorium is Wikisource's community discussion page. Feel free to ask questions or leave comments. You may join any current discussion or start a new one. Project members can often be found in the #wikisource IRC channel webclient. For discussion related to the entire project (not just the English chapter), please discuss at the multilingual Wikisource.

Contents

Announcements[edit]

Note
This section can be used by any person to communicate Wikisource-related and relevant information; it is not restricted. Generally announcements won't have discussion, or it will be minimal, so if a discussion is relevant, often add another section to Other with a link in the announcement to that section.

Proposals[edit]

Possible bot to determine relevant popular wikipedia pages[edit]

I know that there are bots in wikipedia which determine the number of times a given article is viewed per month. The bot(s) involved generate pages like w:Wikipedia:WikiProject Christianity/Popular pages. I wonder if there might be a way to request that bot, or some other bot, generate a similar list of the articles about either authors whose works are in the public domain and/or specific works which are in the public domain. If it were possible, maybe using such a bot to help decide which books to collaborate on here might be helpful. John Carter (talk) 18:12, 25 March 2015 (UTC)

BOT approval requests[edit]

Help[edit]


Correspondence between Gandhi and Tolstoj[edit]

Dear Madam/Sir,

I am looking for facsimiles of letters (hand-)written bei Tolstoj and came upon the above correspondence. Can you advise whom I have to contact in order to purchase such facsimiles?

Thanks for your cooperation and reply.

Best regards from Germany,

Heidi Hacker

Hope you have seen the talk page where link to certain images are provided.--Arjunaraoc (talk) 13:11, 2 February 2015 (UTC)

When to use 'uc' template to transform text to uppercase[edit]

Under what circumstances would it be appropriate to use the {{uc}} template? This template transforms the given text to all capital letters -- for example: {{uc|example}} produces example.

Recently I've been proofreading/validation documents produced on a typewriter that use all uppercase headings, and have been concerned that maybe I've been doing it wrong by just writing the headings in all capitals instead of using the aforementioned template.

Any guidance would be appreciated. (There are similar templates {{lc}} (lowercase) and {{capitalize}} (first letter capitalized) that I similarly do not know the proper circumstances of, but which I've not had the opportunity to use to date.)

Best. -- Mukkakukaku (talk) 04:26, 8 December 2014 (UTC)

When should you use it? Never. Highlight example, copy it, paste it into a text editor: you get "example". Turns out it wasn't upper-case text at all, it was lower-case text masquerading as upper-case text. The template is evil. It should be deleted. Hesperian 06:00, 8 December 2014 (UTC)
Disclaimer: some browsers are smart enough to push upper-case to the clipboard, so your mileage on my copy-paste demo may vary. I think my point stands. Hesperian 06:01, 8 December 2014 (UTC)
How does the 'evil' argument not equally condemn {{sc}}? {{sc|Example}} (Example) cuts/pastes as "Example", but it looks like E{{x-smaller|XAMPLE}} (EXAMPLE) which cut/pastes as "EXAMPLE". Not too sure where you are going with this argument? 121.218.57.230 06:50, 8 December 2014 (UTC)
"Example" is obviously styled, and I am comfortable with it decomposing to "Example" when that styling is removed. "EXAMPLE" has the appearance of unstyled text, it can easily be rendered using unstyled text, and I am not comfortable with it actually being completely different text with a surreptitious styling applied to it. Hesperian 00:28, 9 December 2014 (UTC)
Not precisely a defence of this template, but there is a good case to be made for using the parser-function equivalent {{uc:example}} — which produces EXAMPLE — which does screen-scrape correctly and reliably as [EXAMPLE] — and that usage is within templates which may need to compare two strings in a case-insensitive fashion. Simply consistently uppercase/lowercase/capitalise both quantities to be compared (say) in a {{#ifeq:}} test.

For reference these parser function/magic keywords exist and produce results as shown:

  1. {{uc:eXaMpLe}}: produces EXAMPLE;
  2. {{lc:eXaMpLe}}: produces example;
  3. {{lcfirst:eXaMpLe}}: produces eXaMpLe; and
  4. {{ucfirst:eXaMpLe}}: produces EXaMpLe.
N.B. Functions 3 & 4 affect only the initial letter of the string, so to produce sentence capitalisation something akin to {{ucfirst:{{lc:eXaMpLe}}}} would be required (produces: Example as expected.) (Buried in "official" documentation about two-three screenfulls down from here: MW:Help:Magic_words#Formatting) AuFCL (talk) 10:04, 8 December 2014 (UTC)
I don't prescribe to the "never" scenario, I would say "hardly ever". My commentary is that I use it with some newspaper articles where the capitalisation has been made by the sub-editor, not the author. I use it as newspaper articles when they come back from search engines can look butt ugly. So I get the presentation form for the article, though the text as for a search engine. — billinghurst sDrewth 12:18, 8 December 2014 (UTC)
Somewhere I got the idea that {{uc}} is for text that is uppercased for emphasis or as a matter of typographical style, but would ordinarily be mixed-case. I used it on this page to produce text that appears to almost exactly match the source (except for vertical spacing, because vertical whitespace doesn't really make sense in hypertext), but, if copied and pasted, comes out in normal title case. In short, I use it to preserve the appearance while allowing the text itself to conform to modern case usage if the markup is stripped, since plain text is easier to read in mixed case. As long as headings would remain distinguishable in some way, I support using {{uc}} with mixed-case text. (I have some older reading devices that can handle plain text but not HTML markup, so I tend to think about these kinds of issues more than most, I guess.) Pathore (talk) 02:04, 3 February 2015 (UTC)
I also don't agree with "never". The template uses CSS for styling which is completely appropriate. In other words, the correct capitalization is stored "behind" the styling. This important for parsing the text, such as is done by search engine spiders. I tend to change all caps from books to the correct case and style it with the template (which uses CSS) as needed. However, I notice many books where the all-caps text has been copied and I don't agree with this as style and content should be separated as much as possible. (And yes, I know that it can't be separated as much as it should be on WS.) The Haz talk 01:13, 24 February 2015 (UTC)
Nevertheless, all that laziness in applying 'template approaches' such as this uppercase based one today will just make more maintenance work for somebody else one day in the future.

Hesp was right; Never apply is best because today's specs regarding this nuance are dumb at best -- the future CSS3 implementations will allow for more flexibility in better defined scenarios for titling-case instances, section-label instances and similiar. -- George Orwell III (talk) 02:30, 24 February 2015 (UTC)

After Action Report 770th FA[edit]

I have in my possession, what I believe to be an authenic after action report, dated 21 August - 30 Sept 1944. My father, Robert F Williams was a T/Sgt in said unit.

I am wondering what, if anything, this report is worth.

If anyone sees this...please email me at:

    umberado@comcast.net

David Williams 23 January 2015 (UTC)


May I submit paid old news article?[edit]

There is an old news article "FORMOSA KILLINGS ARE PUT AT 10,000" on New York Times. It was reported in 1947, 68 years ago. New York Times now charges some money for people to download the PDF version of the article. (Although a search on Google shows many results which allow people to read the full article, like this one.)

If I purchase the PDF version of the article, may I create a page for it?--Matt Smith (talk) 10:23, 23 January 2015 (UTC)

We cannot publish it on Wikisource. The material has to be prior to 1923 to be in the public domain. That is why they are selling it.— Ineuw talk 10:34, 23 January 2015 (UTC)
Thank you for the information. Does the "be prior to 1923" rule only apply on news articles? I see the "posterior to 1923" United Nations Convention on the Law of the Sea. And on Chinese Wikisource I also found an old Chinese news article, which was reported in 1951.--Matt Smith (talk) 11:13, 23 January 2015 (UTC)
Actually, I want to translate that article into Chinese and then submit it onto Chinese Wikisource. English Wikisource isn't the place I want to submit an article to.--Matt Smith (talk) 11:33, 23 January 2015 (UTC)


Articles that are in the public domain in the US may be freely hosted here. "The material has to be prior to 1923" is not quite correct: texts published before 1923 are almost definitely in the public domain in the US; texts published after 1923 probably are not. I don't know about this particular article.

But even if it turns out that we can host the text, that does not imply that you have the right to post the text. If in purchasing an article you enter into a contract that forbids you to redistribute it, you may find yourself in breach of contract if you post it here.

Hesperian 11:38, 23 January 2015 (UTC)

Thank you. It looks like the article isn't in the public domain otherwise New York Times wouldn't sell it.--Matt Smith (talk) 12:51, 23 January 2015 (UTC)
My last two cents. It would be best to look up and read Wikipedia and Wikimedia commons articles about the subject. Both have extensive copyright information. — Ineuw talk 16:41, 23 January 2015 (UTC)
I really hope the article can be posted on Wikisource and your advice is inspiring. However, since New York Times is still selling the article, is it possible that there is some copyright rules allow the article to be posted on Wikisource?--Matt Smith (talk) 02:06, 24 January 2015 (UTC)
People selling something doesn't mean much one way or the other. But the New York Times was renewed basically from 1923 on, so most of the newspaper (sans advertising) is still under copyright.--Prosfilaes (talk) 10:56, 3 February 2015 (UTC)

Text on top of image[edit]

I've forgotten how to. Page:Child-life in Japan and Japanese child stories (Ayrton, Matilida Chaplin. , 1901).djvu/46. Moondyne (talk) 13:53, 29 January 2015 (UTC)

You may have a try with {{overfloat image}}. Hrishikes (talk) 14:05, 29 January 2015 (UTC)
Ahah! Thanks, and wish me luck! Moondyne (talk) 14:23, 29 January 2015 (UTC)
I cannot seem to remove whitespace below the image. Is there a trick? Moondyne (talk) 15:12, 29 January 2015 (UTC)
Yes check.svg Done Hrishikes (talk) 15:44, 29 January 2015 (UTC)
Hrishikes, that remedy is absolutely fascinating! —Maury (talk) 22:56, 1 February 2015 (UTC)

Is this regular or italicized Greek?[edit]

Having no basic knowledge whatsoever about Greek characters, I would be interested in any advice whether the Greek characters in A Dictionary of Christian biography and literature, like at Page:Dictionary of Christian Biography and Literature (1911).djvu/20, are "regular" Greek characters or italicized. John Carter (talk) 16:40, 29 January 2015 (UTC)

"regular". The preference here is to avoid italicising non-Roman text because the fonts aren't designed with italic options. I usually wrap Greek characters in the {{Greek}} template because the fonts in there are more easily read, particularly the diacritics. Beeswaxcandle (talk) 16:47, 29 January 2015 (UTC)
In this case, it's "regular" text, but the characters are Ancient Greek (not the modern language). The ancient form of the language includes characters and diacritics not present in the modern language. The standard template for enclosing Ancient Greek text on most MW projects is {{polytonic}}. --EncycloPetey (talk) 20:27, 1 February 2015 (UTC)

The Dream of Life poem[edit]

I'm not sure how the quote at the start of the poem is formatted. Besides, I changed to a new computer and a new, wide monitor quite recently. --Lo Ximiendo (talk) 16:38, 31 January 2015 (UTC)

I gave it a shot; let me know if it is acceptable! Londonjackbooks (talk) 17:23, 31 January 2015 (UTC)
Thanks for your help along with the help of Mpaa. --Lo Ximiendo (talk) 20:28, 31 January 2015 (UTC)
No problem :) Londonjackbooks (talk) 22:35, 31 January 2015 (UTC)

Can legal papers be published?[edit]

Can legal papers which were submitted to Supreme Court of the United States in 2009 be published? The papers are someone's affidavit for testifying the case.--Matt Smith (talk) 13:39, 1 February 2015 (UTC)

@麥特-斯密斯: We usually don't publish something for the first time here. Can you tell us more about what the documents are? —Justin (koavf)TCM 15:11, 2 February 2015 (UTC)
The documents (affidavit) were wrote by "former President of the ROC government in exile". They were used in a lawsuit against the United States government. Please see this reporte.--Matt Smith (talk) 07:42, 3 February 2015 (UTC)
@麥特-斯密斯: Probably. I would imagine that anything submitted to a court would by definition be public record and if it were written by an ROC president, the author would be notable. Let me know how I can help. —Justin (koavf)TCM 05:03, 4 February 2015 (UTC)
That sounds good. What's the appropriate way for me to hand the documents to Wikisource.org?--Matt Smith (talk) 05:56, 4 February 2015 (UTC)
@麥特-斯密斯: Generally, if it was originally published in print, then there will be a digital scan of it that is uploaded to our sister site Commons. This is not strictly necessary but it is useful sometimes. If you have digital copies, then they can be virtually copied and pasted here. —Justin (koavf)TCM 06:01, 4 February 2015 (UTC)
I see. I found the pdf version on the personal website of the petitioner and the image version on a Chinese website. Are those versions okay for uploading to Commons?--Matt Smith (talk) 06:36, 4 February 2015 (UTC)

@麥特-斯密斯: Definitely. Can you tell if the two are identical? PDFs are nice because we can use software here to scan the document and reproduce the text. —Justin (koavf)TCM 06:38, 4 February 2015 (UTC)

Yes, I believe they are identical because the font and the signature on the the pdf version look identical to the image version. Obviously the image version was created by scanning the pdf version. Or rather, the pdf version was created by scanning the image version. Anyway, they look identical. --Matt Smith (talk) 06:54, 4 February 2015 (UTC)
From my reading of the United States Code/Title 17/Chapter 1 I believe that the text you are proposing to add is copyright and not in the public domain. While court decisions are not copyright, there is no explicit exception made for court proceedings or for the documents submitted to courts. For us to host this text, there will need to be an OTRS ticket showing that the author gives permission to Wikimedia to host. Beeswaxcandle (talk) 07:25, 4 February 2015 (UTC)
I understand. Then I might need to talk to the petitioner to ask if he is willing to give permissions to Wikimedia. Since I myself and probably the petitioner have no clue on how OTRS and related copyright rules work, what should I inform the petitioner? And what effects will happen to his documents after the text is hosted on Wikimedia? He may need to know these things before making the decision.--Matt Smith (talk) 07:51, 4 February 2015 (UTC)
The best thing I can do is to point you to Commons:OTRS. The process is explained there. Beeswaxcandle (talk) 08:09, 4 February 2015 (UTC)
Thanks. I will read it.--Matt Smith (talk) 08:45, 4 February 2015 (UTC)

Why no automatic ocr text layer ?[edit]

Yesterday I uploaded from IA to Commons the book and then at wikisource I created Index:Irish_Emigration_and_The_Tenure_of_Land_in_Ireland.djvu after spending considerable time creating the entry Author:Frederick Temple Blackwood. Now the index Progress is "Index_-_Text_Layer_Requested" Does this mean I should do something to get the text layer ?

This is the second book I added to wikisource. Previously for Index:Annalsoffaminein00nich.djvu it seemed the ocr layer was automatically present. I have read soooooo many help pages but this situation does not seem to be discussed. Thanks.Robin2014 (talk) 16:25, 1 February 2015 (UTC)

The automatic text layer is generated at IA and incorporated into the DjVu file there. If the file is missing its OCR text layer, then we won't have it either. We don't generate this layer ourselves. The IA procedures change over the years, and some of their DjVu files are missing one or more of the layers that are now standard.
This question comes up often. Could someone well-versed in the issue (and how to resolve it) start a page with guidance? --EncycloPetey (talk) 20:24, 1 February 2015 (UTC)
This djvu file has a text layer. I do not know why the extension does not show it. My opinion is that there is some problems with the text format and the extension does not recognize it. Might be wrong, 2nd opinion welcome.--Mpaa (talk) 20:32, 1 February 2015 (UTC)
It was derived back in 2008 for one thing. Chances are there are structural errors affecting the coordinate mapping or the outline labeling. One sure way to check is to extract the text-layer and try merging it back in. If it extracts but fails upon the [re]merge, you should get some sort of error message indicating why it failed (then we go troubleshooting based on that next). -- George Orwell III (talk) 20:40, 1 February 2015 (UTC)
Like I said....

┌────────────────┘

C:\Program Files (x86)\DjVuLibre>djvused duff.djvu -e 'output-txt' > duff.txt -u

C:\Program Files (x86)\DjVuLibre>djvused duff.djvu -f duff.txt -u -s

*** Syntax error in txt data: illegal zone token 'page',
        near ' 0 0 1720 2994'
*** (..\..\..\tools\djvused.cpp:380)

... it seems the OCR routine at the time could not handle complex tables that take up almost an entire page so it "thought" the table(s) were pages themselves -->
--> (illegal zone token 'page'). -- George Orwell III (talk) 21:03, 1 February 2015 (UTC)

@Robin2014: I can upload the current text layer by bot. Let me know if you want to try to re-obtain a better text layer from IA or if you want me to go ahead with the current one. You can judge the quality looking at this page.--Mpaa (talk) 22:49, 1 February 2015 (UTC)
.... or you can just hit the OCR button for every page; same result. Doesn't matter if its a lump dump or a page a time; garbage in, garbage out I say. The most straight forward solution is to take the original .PDF and re-derive it by re-uploading to IA using a different identifier (like ' irishemigrationt01duff ' for example). Once the processing is done, you can take the new .djvu and upload it over the bad one on Commons. -- George Orwell III (talk) 00:10, 2 February 2015 (UTC)
Thank you all for the info. All 4 copies of this book on IA were processed in 2008. I decided to take the pdf file and upload it to IA with a slight change in the file name. I'll take a look at it when IA processing is finished and let you know if it seems a better result. Thanks again Robin2014 (talk) 14:24, 2 February 2015 (UTC)
Would simply requesting re-deriving the file on IA possibly solve the problem? Would it be worth considering it a standard practice before uploading the file to Commons to make sure the IA OCR was generated at least within the past couple years? djr13 (talk) 14:54, 2 February 2015 (UTC)
I haven't had much luck using their manager for such things personally but its something worth looking into further - I just have too much on my plate at the moment. I think Nemo bis is the best person to tap for guidance on this (I believe he's an IA admin). -- George Orwell III (talk) 15:30, 2 February 2015 (UTC)

djr13 and Robin2014 pointed me to this discussion. It's certainly appreciated, both by Internet Archive and by users, if a new derivation is made and quality tested with ABBYY 9.0. However, I don't have sufficient privileges to rederive an item which was successfully derived, unless I uploaded it. The recommendation IA gave me is to upload a duplicate with the same metadata and originals but a different identifier: everyone can do this, and is encouraged to.

If you find a pattern where books in language X or in topic Y are considerably better with the new OCR, we can tell IA and they will rederive en masse. If reuploading yourself is too burdensome, I can help, but not on a continuous basis: give me a list of a few dozens/hundreds items of which you want to verify a new OCR, and I will reupload them for you. Ideally we'd then return, to IA, a table listing the tested IDs and what improvements there was. --Nemo 22:27, 2 February 2015 (UTC)

IA has processed the book under the name https://ia802600.us.archive.org/3/items/irishemigrationt01duff. I used URL2Commons to move this .djvu to commons with name "Irish Emigration and The Tenure of Land in Ireland 2.djvu" I could not put this in commons with the original name so I added '2' at the end of the title. Then I created the Index with '2' at the end of the name and now the OCR layer was done automatically. I guess this was the result of the re-derivation of the djvu. Now what does wikisource prefer - should the name without the 2 be deleted ? OR should be name with '2' be used to replace, or update the name without '2' ? Robin2014 (talk) 15:51, 3 February 2015 (UTC)
I reupload it at Commons as a new version of the original one, deleted the Index:*_2.djvu here and nominated the File:*_2.djvu at Commons.--Mpaa (talk) 20:03, 3 February 2015 (UTC)

Whitespace below {{overfloat image}}[edit]

Above leaves a block of whitespace below each image and the examples on the /doc page all seem to show this. Viewing page source shows the anomaly is this

<p><br /></p>
<p><br /></p>
<p><br /></p>
<p><br /></p>

Where does that come from and can it be removed? Moondyne (talk) 02:12, 2 February 2015 (UTC)

Short answer: from inside the template itself. Happy now? 124.183.99.79 04:10, 2 February 2015 (UTC)
Always happy. Cheers. Moondyne (talk) 07:08, 2 February 2015 (UTC)
Good to hear. My earlier note was unreasonably terse. What I should have said is that the comments embedded inside the template were slightly mis-aligned; resulting in issuing one blank line for every possible "item" field (there are nine of these: item1..item9.) This becomes significant when you realise 8 empty lines represents 4 paragraph boundaries... thus your anomaly in a nutshell. Hope this (longer) explanation is clear enough? 124.183.99.79 aka 101.175.138.189 07:24, 2 February 2015 (UTC)
Yes it is. I am much obliged. Moondyne (talk) 23:26, 5 February 2015 (UTC)

Small help with my first text[edit]

Original_Stories_from_Real_Life This needs a few chapter transclusions to be fixed and the table of contents validated. User:John Carter and I have put in most of the work if someone else wants to help polish it off. Thanks! —Justin (koavf)TCM 15:12, 2 February 2015 (UTC)

It looks as though the Google notice also needs to be stripped out. --EncycloPetey (talk) 15:38, 2 February 2015 (UTC)
Extraneous initial pages taken care of. Hrishikes (talk) 16:20, 2 February 2015 (UTC)
EncycloPetey, how would *you* strip out the Google watermarks on any file, .pdf or .djvu? —Maury (talk) 16:18, 3 February 2015 (UTC)

Charinsert bar disappeared[edit]

The title says it all. Can anyone shed some light on this? — Ineuw talk 21:24, 3 February 2015 (UTC)

Getting back to this issue, is there anyone who knows how to fix this? I am like a fish out of water, a boat without oars, and just drifting here. — Ineuw talk 11:37, 4 February 2015 (UTC)
Just helping myself - turned the gadget feature off and then on, and now it's back. Thanks ineuw, you're great.— Ineuw talk
Charinsert disappeared again - this time disabling & enabling didn't help. Can someone help to resolve this please? — Ineuw talk 22:28, 4 February 2015 (UTC)
Gone for me too (Chrome)... But it appears in IE. Anxious to get it back! Londonjackbooks (talk) 03:35, 5 February 2015 (UTC)
Still there for me in Firefox and Safari, but strangely it's at the top of the edit window in Safari while at the bottom of the edit window in Firefox. --EncycloPetey (talk) 03:39, 5 February 2015 (UTC)
When it was appearing in Chrome, it would normally appear at the top, but would occassionally appear at the bottom. But at least it was there! Londonjackbooks (talk) 03:52, 5 February 2015 (UTC)
It's there for me in chrome (android), I have just now done some edit with it. Hrishikes (talk) 05:22, 5 February 2015 (UTC)
I guess the lights went out and GO3 is not around to blame. It's not appearing in any of my browser toys in any of the OS's.— Ineuw talk 05:24, 5 February 2015 (UTC)
I've lost it as well (it seemed to appear sporadically for a short time, then disappear entirely). The code behind it is MediaWiki:Gadget-charinsert-core.js. That was last edited on 23 Jan by @Krinkle:, which doesn't match up with problems occurring over a week later (besides which he appears to be a developer -- I'm afraid I don't keep up with that side of things -- and should know what he's doing). I can't see any other obvious edit on any related page that might be causing this. - AdamBMorgan (talk) 14:19, 6 February 2015 (UTC)
Filed a bug report on Phabricator.— Ineuw talk 06:16, 7 February 2015 (UTC)
All appears normal for me ... monobook, old toolbar above, and the edittools below. — billinghurst sDrewth 06:24, 7 February 2015 (UTC)
Ah, ha! It's not gone for me either, but I'm still using monobook as well. Has it gone for anyone who is using monobook? If not, then that needs to be part of the bug report. Beeswaxcandle (talk) 06:37, 7 February 2015 (UTC)
Still gone in Chrome; hopefully the bug ticket will get someone to look into why the problem exists in some browsers and not in others. May need to be specific in the report. Londonjackbooks (talk) 06:29, 7 February 2015 (UTC)

┌───────────────────────────────────────┘
I was using the modern skin, but switched to another to test it and there was no change. At my end it doesn't seem to be skin or browser related because it doesn't work in any of the browsers. Perhaps it may be a combination of selected gadgets? I also tried to place the code in the skin related .js folder, but that didn't help. Currently it's installed in the "Shared CSS/JavaScript for all skins". — Ineuw talk 06:53, 7 February 2015 (UTC)

What about now that I removed the Jan 23 Krinkle toolbar bit addition and dependency changes? -- George Orwell III (talk) 21:19, 7 February 2015 (UTC)
I was posting the this when GO3's post collided with mine: "This problem also popped up in other wikis. I disabled all gadgets as it was suggested, but nothing happened since my last post until now. The Charinsert has returned. Whether this is permanent or not, I don't know, but I am very suspicious that the return of the Charinsert coincides with the reapperance of User:George Orwell III." :-).— Ineuw talk 21:26, 7 February 2015 (UTC)
The problem was 2 weeks off for Developer Reach-Around Camp on top of the normal 1 week wait for a new 1.25wmf release to come down to us making for a total of 3 weeks between developer "action" and developer "result". The change this past tues resulted in 3 weeks of good intentions going bad it seems (I can't remember what took place 3 days ago so its not hard to believe forgetting something 3 weeks old taking place). -- George Orwell III (talk) 21:33, 7 February 2015 (UTC)
I appreciate your explanation, but it doesn't mean that I accept it. For all I know you were off on a quest for better goulash.— Ineuw talk 21:43, 7 February 2015 (UTC)
Sorry, I forgot to advise the community that the toolbar is back again (for me). — Ineuw talk 21:52, 7 February 2015 (UTC)
Back, and functioning. Thanks, Londonjackbooks (talk) 23:53, 7 February 2015 (UTC)

Missing features in non-English Wikisource?[edit]

I'm considering putting some Latin texts on Wikisource. However, the Latin Wikisource seems to have much fewer features than the English Wikisource. For example, margin notes, verse numbers, TOC options--all these are missing, and they are important for these texts. I assume Latin texts cannot be hosted on the English Wikisource. Is there any solution?— Fps.vogel (talk) 19:00, 3 February 2015 (UTC)

One I can think of is to copying the templates to the Latin site.— Ineuw talk 00:49, 4 February 2015 (UTC)
Got it, thanks!— Fps.vogel (talk) 15:19, 6 February 2015 (UTC)

Corruption in {{overfloat image}}?[edit]

Suddenly I noticed today some extra curly braces in a page proofread by me with the help of {{overfloat image}} and already validated. Has someone tinkered with the template? Hrishikes (talk) 14:25, 5 February 2015 (UTC)

You are right. The attempt to address Whitespace below {{overfloat image}} above introduced this error. @AdamBMorgan: has also been working on this template; and as my solution is quite different to his (he had only addressed the issue insofar as item1 was affected) I shall let him decide which approach to crystallise upon. In short, please check back from time-to-time to ensure this thing is (still) working.
Perhaps excessively ramming the point home: two-edit change differences. This is important (possibly only to ABM above) but probably not interesting. 110.147.190.65 21:23, 5 February 2015 (UTC)
I had to stop to do something else. I see what the problem is now, but I went ahead with just replacing the table segments with straight HTML instead of wikicode. HTML has explicit close tags that don't require interpretation. Hopefully this will also prevent future errors as well as the current bug. Besides, all the wikicode does is get translated into HTML anyway. I've taken the time to clear up the code as I found it difficult to interpret as it was; a little more spacing and structuring (with commented out whitespace) should help there. - AdamBMorgan (talk) 00:00, 6 February 2015 (UTC)

Index:Bradshaw's Monthly (XVI).djvu[edit]

Finally completed, but I'd like someone to advise on how to tweak the tables slightly.ShakespeareFan00 (talk) 17:32, 6 February 2015 (UTC)

The only reasonable tweaks I can see at a glance might be replacing the blanks and some minor formatting bits. Where you currently have
|{{ts|ar}}|.
|{{ts|ar|br}}|.
You could try |colspan="2" {{ts|ac|br}}|. ., which gives a result closer to the original. In general, collecting the pairs of numbers on single lines also makes the wikitext easier to work with. I've just done a bunch of this on Page:Bradshaw's Monthly (XVI).djvu/8. Was this a series that ran consistently enough and long enough for specialized table formatting templates to be worthwhile or where the tables radically different in each edition? Pathore (talk) 04:16, 7 February 2015 (UTC)

| "Helper" templates was how i got this proofread. the issue you mention above is a sideeffect, which ca be easily cleaned up. ShakespeareFan00 (talk) 16:31, 7 February 2015 (UTC)

OCR button has disappeared[edit]

The OCR button is gone from the toolbar. This is a pretty serious problem for working with source files that contain no text layer. --EncycloPetey (talk) 17:46, 7 February 2015 (UTC)

Present in chrome in Windows 7 in my laptop. Absent in IE and Firefox. Hrishikes (talk) 18:20, 7 February 2015 (UTC)
OCR button not present in Safari on iMac. --kathleen wright5 (talk) 22:18, 7 February 2015 (UTC)

┌────────────────┘
Please check again -- any change now? -- George Orwell III (talk) 03:39, 8 February 2015 (UTC)

Now it has disappeared from Chrome (Windows 7) as well. Absent in IE and Firefox as before. Present in Chrome (android). Appears in Chrome (Windows 7) after logging out. Appears in Firefox after logging in. Hrishikes (talk) 03:56, 8 February 2015 (UTC)
Still no OCR button in Firefox for MacOS. It showed up briefly in Safari (MacOS), but when I logged out and then back in, it disappeared again in Safari. --EncycloPetey (talk) 05:24, 8 February 2015 (UTC)

┌────────────────┘

@Kathleen.wright5, EncycloPetey, Hrishikes: I'm convinced this latest drop-out of the OCR button has to do with last Tuesday's core 1.25wmf15 update and it's "effect on the way things [gadgets] are loaded when they are labeled as site-wide defaults" (I don't know how else to phrase it never mind identify what it was specifically). Soon afterward, I'm pretty sure the combination of CharInsert being a site-wide default, the OCR button being a site-wide default and the ProofreadPage toolbar extension in general are now all "competing to load at the same point in time & into the same memory".

As a work-around that produces the most consistent results for all three components [here on my lonely IE11 Win Edge setup], the OCR button gadget will no longer be a site-wide default for the time being (maybe forever?) -i.e. >>>> YOU MUST ENABLE THE OCR BUTTON GADGET MANUALLY IN YOUR USER PREFERENCES NOW <<<<

Once manually enabled AND given some time for the change to actually work its way through the existing cache or caching for each user, the OCR button should then reliably generate for you in each and every Page: edit session you initiate and leave the PR extension and the CharInsert toolbar to their own "functioning states" at the same time.

Please post back with your observations and/or results. -- George Orwell III (talk) 05:50, 8 February 2015 (UTC)

Back in Chrome after manual preference. Thanks. Hrishikes (talk) 06:06, 8 February 2015 (UTC)
That OCR button must be on the newest editor version. I have *never* had to use the OCR button on this older editor. Always try to use .djvu files. —Maury (talk) 07:56, 8 February 2015 (UTC)
@William Maury Morris II: The problem is that some DjVu files do not have a text layer, and then you either have to redo the file at IA, or else use OCR. --EncycloPetey (talk) 16:33, 8 February 2015 (UTC)
Still not seeing the OCR button in Firefox or Safari (MacOS), even after logging out/in and patient waiting. I set my Preferences for the Gadget shortly after GO3 posted. --EncycloPetey (talk) 02:17, 9 February 2015 (UTC)
@EncycloPetey: Interesting. I have been blessed, I suppose, because the .djvu files I upload all have text layers and therefore I never had to use our OCR button. Still, my OCR is still exists to use. I see no problem in anyone having to redo a file at IA. I learned from Ineuw, or however he spells his alias here, how and where to get and/or derive .djvu files and to avoid .PDF files when possible. Heaps of blesssings upon him for it! I recall he placed detailed instructions on wikisource! Too, I think George Orwell III has explained it. I use only Firefox 35.0.1 until it is upgraded and never do I have to use the OCR button. —Maury (talk) 02:37, 9 February 2015 (UTC)

┌──────┘
@George Orwell III: I should create a table for all the browsers in various OS's. I tested the OCR in the following browsers using Ineuw for the advanced toolbar and my public account configured for the legacy toolbar.

  • In both Windows and Apple - Logging out and then in works if not on the first try then the second.
Advanced toolbar
  • In Windows and Apple - It always works with the advanced toolbar in all browsers. This includes, Firefox, IE 11, Chrome, Opera and Safari.
Legacy toolbar
  • In Windows - It works in all browsers except in Opera, where logging out and in brought it up on the second try.
  • In Apple - Always works in Firefox but not in Safari. Logging out and in brought it up on the second try.

I also tested if it works for each of the above. On one occasion in Safari, it showed up but didn't work on the first try. If anyone has a problem with Linux, I can test it in Xubuntu 14.04 with Firefox, Opera, and perhaps Chrome (don't remember if I installed it). I really hope it helps. — Ineuw talk 04:52, 10 February 2015 (UTC)

I tested both Firefox and Chromium in Linux and the results were that with the advanced toolbar OCR always shows up and works. With the legacy toolbar the icon was missing on the first edit in both browsers. Purging the page (I use the UTC clock purge), it shows up. Perhaps the editors with a problem should check their preferences that only one toolbar is selected.— Ineuw talk 17:39, 10 February 2015 (UTC)
Indeed, that was the problem in my case; I had more than one version of the editing toolbar selected (no idea why or how). Correcting this issue to select just one of them corrected the problem. Thanks to GO3 for helping me figure that out. --EncycloPetey (talk) 02:54, 11 February 2015 (UTC)

Sir Arthur Stanley Eddington[edit]

Hellow!
Works of Sir Arthur Stanley Eddington are in the public domain now. So I've made Index:Eddington A. Space, Time and Gravitation. 1920.djvu. Please proofread it. --Максим Пе (talk) 13:43, 9 February 2015 (UTC)

A note to editors: The title of the work is Space Time and Gravitation. The comma in the name of the DjVu file is erroneous (it will not affect editing the work). The work is on the interplay of (a) space-time and (b) gravitation, and is not discussing three separate topics. However, it would be nice if someone with the use of a bot could rename the file at Commons and move the few pages created here so far. --EncycloPetey (talk) 19:42, 9 February 2015 (UTC)
Is that sufficient grounds for any of us to get someone at Commons to rename the file or does Максим Пе (talkcontribs) need to put in the rename request (as "uploader requested") at Commons? There are only 16 pages done so far, few enough that I'll do the moves manually, if the rename is done before too many more pages are added. Also, what is going on with that Table of Contents on the Index page? I see it overlapping with the index fields. Pathore (talk) 22:48, 9 February 2015 (UTC)
If the pages here are moved, then I can do the renaming at Commons afterwards. A move/rename at Commons is not so big a thing now as it used to be.
The table of Contents on the Index page is a temporary version inserted manually, and it can be replaced once the Contents in the source file is proofread. --EncycloPetey (talk) 00:42, 10 February 2015 (UTC)
@EncycloPetey:I've just finished doing the moves here; go ahead and rename the file on Commons. It looks like the page list will be broken until the rename goes through there. Pathore (talk) 03:50, 10 February 2015 (UTC)
I've also found the actual cause of the page layout problem: the "Cover image" field is supposed to be a page number, it seems, but was an actual [[File:]] link instead. Now that I've fixed it, the cover image is also a red link until the Commons rename goes through. Pathore (talk) 04:36, 10 February 2015 (UTC)
Move I'm not sure why this took so long but I happened to see that this file needed renaming because I was alerted through Special:Notifications when User:Максим Пе thanked me. I moved it on Commons and then came back here and saw all this. I'm not that interested in proofreading this text but I'll help validate if someone pings me on it. —Justin (koavf)TCM 05:08, 10 February 2015 (UTC)

┌────────────────┘
This is a bit of a mess. Unfortunately the proper name on the Commons should have been File:Space, Time and Gravitation (1920).djvu which currently is a redirect to the bad file name. I can move files but this requires a delete which I can't do. One must ask for help and explain that the redirect should be the right file name.— Ineuw talk 05:11, 10 February 2015 (UTC)

Do the pages here on Wikisource need to be moved again? Also, there is no comma in the title, so the correct name would be File:Space Time and Gravitation (1920).djvu, which does not exist at the time of this writing. Pathore (talk) 05:16, 10 February 2015 (UTC)
Move I don't understand the problem: the index is fine now and synced with Commons. I've since validated and edited pages to this text. —Justin (koavf)TCM 05:19, 10 February 2015 (UTC)
Now I'm getting confused. As I see it, the file was uploaded under the name "Eddington A. Space, Time and Gravitation. 1920.djvu". It was pointed out that there is no comma in the title of the work, so I moved all of the pages to remove the comma, but I can't move files at Commons, so I left the note in the index. Now it is being pointed out that "Eddington A. Space Time and Gravitation. 1920.djvu" doesn't follow the standard naming convention in some other way that I am just now learning about. 16 pages isn't too many to move by hand, but if the work gets mostly proofread before we decide on its file name, things get messier. Pathore (talk) 05:27, 10 February 2015 (UTC)
Everything fine as it now stands. Both our files and the file on Commons have been correctly renamed, as they should be. Everything is fine. It only looked like a mess during the brief period after we had renamed our local files, but the source file on Commons had not yet been moved. Once everything was moved, the mess was resolved. --EncycloPetey (talk) 15:06, 10 February 2015 (UTC)
@Ineuw: There is no reason to add a comma to the book's title. It never had one. Einstein's concept of spacetime can be written with a blank space (space time), a hyphen (space-time), or be written as a single word (spacetime), but it cannot be written with a comma in the middle of the term. That would be like saying you "rode a school, bus" instead of saying "rode a school bus". It's nonsense to insert a comma into the middle of a term. --EncycloPetey (talk) 15:21, 10 February 2015 (UTC)


OMG! I think you are missed.
First of all filenames are very free at Commons. For example File:LorentzStatement1920.djvu => Index:LorentzStatement1920.djvu => The Einstein Theory of Relativity
I've put filename "Space, Time and Gravitation" with comma because that was in sourceInternet Archive. Also please look at Project Gutenberg: Space, Time and Gravitation by Sir Arthur Stanley Eddington and Amazon: Space, Time and Gravitation: An Outline of the General Relativity Theory (Classic Reprint)
At last I don't know why there is no comma in first edition title. Really :-) May it be misprint? --Максим Пе (talk) 14:40, 10 February 2015 (UTC)

No, the comma was not in the source (the original book), it was an arbitrary file name choice at IA for their file name. Nor is it necessary for us to keep the IA name when uploading to Commons. There is no comma is the title of the work, nor should there be. Presumably what happened is that someone unfamiliar with the concept of spacetime inserted a comma when uploading to IA. The IA file name was an error; the title page of the book is correct. There should never a comma placed in the middle of a noun phrase like "space time".
The reason that Gutenberg and Google also contain the comma is that all three come from the same mistranscribed source error. If you look at the text itself in the copies, there is no comma on the book's title page, only in the name under which they have chosen to list the book. So, again, it is a file name error made by people unfamiliar with the concept of spacetime. --EncycloPetey (talk) 15:03, 10 February 2015 (UTC)
Well, look at one of the first translations: Espace, Temps et Gravitation (French 1921). Also notice Eddington wasn't used term "spacetime" or "space time", but always "space and time". Can't you believe in misprint till now? Максим Пе (talk) 16:21, 10 February 2015 (UTC)
A translation is a different work, and tells us nothing about the original that is useful in this regard. I have seen many translations of titles that were mangled in translation. --EncycloPetey (talk) 02:50, 11 February 2015 (UTC)
Indeed, space-time has no comma; but this consists of space and time and this classic book is one of the first works to describe the concept to the people. This is a book about the inter-relation of space, time and gravitation, where the components together make spacetime and its warping makes gravitation. The comma is present in the title; see the front cover here and here. Hrishikes (talk) 16:19, 10 February 2015 (UTC)
@Hrishikes: Talk with any author, and you'll find that it's the publisher that creates and designs the cover, not the author. Many times the title on the cover of a book will not match the author's desires, or even the title page within the same volume. For this reason, librarians rely upon the title page version of the title, and not the title printed on the cover of a book. --EncycloPetey (talk) 02:50, 11 February 2015 (UTC)
I am fine with a comma, or no comma in the title. Please notice the comma in this post.— Ineuw talk 17:29, 10 February 2015 (UTC)
On the one hand, the work itself consistently uses no comma in its title on both djvu pages 7 and 11. The text itself seems to consistently refer to "time and space"; a search for "space time" gives hits only on the title pages and "space, time" matches only within the phrase "views of space, time, or force" on djvu page 196. Djvu page 13 has a call number—QC6 E42— written on it. A search for this call number in the UC Berkeley online catalog retrieves records listing the title as "Space, time and gravitation : an outline of the general relativity theory". On Wikisource, do we use the title from the work's own title page, or do we use the title as used by actual librarians? Do I need to undo all those moves? Pathore (talk) 23:52, 10 February 2015 (UTC)
We generally follow the title give on the work's own title page, and not what other people have called the work. --EncycloPetey (talk) 02:50, 11 February 2015 (UTC)
There is another angle too. This work is universally referred to with a comma, as can be easily verified with a simple Google search. Here at WS, a work should be placed under its commonly attributed name, so that it is visible in the first page of a Google search when potential readers search it. This should have primacy, not erudite discussion about what the author might have desired. The author is not going to come back from where he is and read it at WS, lay people like we commoners are going to read it; if people have difficulty in finding it, then what is the purpose of putting it here? Hrishikes (talk) 04:08, 11 February 2015 (UTC)
So we will definitely need a redirect, then. Or I'll move it again if three other people can agree on its proper name. Pathore (talk) 06:18, 11 February 2015 (UTC)
There will never be a need to move it again. At most, we can use a redirect, but I suspect that internet search functions can cope with the presence or absence of a comma. --EncycloPetey (talk) 14:32, 11 February 2015 (UTC)

┌───────────────────┘
I think, at this point of the discussion, opinion of other editors are required. So I shall summarize the case, with additional points.

  1. The name of the work, as shown in the front cover of 1987 Cambridge edition as part of Cambridge Science Classics as well as 2011 edition shown by Google Books, has a comma between space and time.
  2. The title page of the version added here does not have this comma.
  3. The version present here is from Internet Archive, which shows the comma.
  4. The IA version is originally from UC Berkeley, whose online catalogue shows the comma.
  5. User:Максим Пе has pointed out that Gutenberg, Amazon and a French translation, all have the comma.
  6. EncycloPetey has opined that:
    1. Space time does not require a comma in between, being a single concept.
      1. Pathore has pointed out that the term space time, without a comma, is present only on the title pages of the work, and not in the main body.
    2. The title page is to be preferred, because the front cover is publisher-designed.
      1. Here I submit additional evidence, to show that the title page is also publisher-designed, at least it used to be at the period and place of this work's publication. At Index:The King of the Dark Chamber.djvu, a reputed publisher like Macmillan made wrong attribution of the translator on the title page. This got corrected only in later editions on written protest from the original author. This episode has been discussed by academicians in scholarly works as found here and here. Therefore, the front cover and title page are both publisher-designed and carry equal value as regards the name of the work.
  7. This is not only about the matter of a comma or only this work. A general policy is required about how to decide the title if there is discrepancy between front cover and title page. Is opinion of other websites, including online catalogues of university libraries and book-sellers, to be given weightage? Is the subject matter of the work, as detailed in the main body, to be given weightage? I sincerely request other editors to offer their esteemed opinion.

Hrishikes (talk) 04:23, 12 February 2015 (UTC)

  • Good lord.. really?
  1. Nobody comes here nor to Commons for a source file - they most likely do the same thing we do - rip it from Internet Archive, HathiTrust, or Google Books ∴ the only thing that contributors should be concerned about for the title they select for a source file being uploaded is that it a.) makes rationale and logistical sense to other potential wiki___ contributors and/or projects more so than any academic consideration(s) followed by b.) the fact at some point and time in existing history, even the presence of the simplest of punctuation gave the "rendering system" enough reason not to render at all. This is a case where the less [punctuation] one applies the less likely [punctuation] will ever cause rendering "issues" in the future.

    Plus its not like the final proofread & validated text is [re]inserted into these source files anyway - if somebody downloads or open one they get the same garbage we did on the day when we first started PR'ing the work.

  2. The same premise outlined above generally applies to the Index: title. Nobody outside us familiar clowns knows or cares about the Index: name or namespace -- especially IF everything has been "done", validated and presented in the main namespace.
  3. The main namespace root page's title is the most important (both internally and externally). Nobody stopping by here gives a sh!t about anything else -- if landing on the mainspace is hindered by "poor naming practices" in reality, I assure you the traffic will go to the places where we ripped the source file from to begin with instead.
-- George Orwell III (talk) 04:46, 12 February 2015 (UTC)
  • @Hrishikes: (Edit conflict) Some of the assertions made above are misleading or at least incomplete in their summary. Specific instances:
  1. The IA version does NOT contain the comma. The comma appears only in their filename. The comma appears nowhere in their scan of the book.
  2. Several instances listed above are not independent of each other, and therefore cannot be counted as independent lines of reasoning. To whit, if the IA version comes from UC Berkeley, then the comma in the IA filename is likely the result of it being present in the UCB catalogue and not of an independent decision.
  3. Only a single title page is mentioned, but it has been pointed illustrated that even an edition with a comma on the cover does not include that comma on the title page.
The assertion of a single instance where a mistake was made in a publication on the title page does not demonstrate that it is of equal weight with the cover. The internal pages of books are routinely sent to authors for verification and proofreading. The cover is not. When books are reprinted, the internal contents are often reproduced as previously published (though not always), but the cover may change completely with a new edition. A single instance where a publisher misattributed the translator of a work does not alter that.
What is also absent from the summary above is that most the discussion has been over our internal filename, which may or may not be reflected in the title we use for the user-accessible version of the work. We could upload a file to Commons as File:Fred.dvju, use Index:Fred.djvu locally, create Page:Fred.djvu, and still title that work Axioms Outlined by Artistotle. Our internal filename and public work title do not have to match, nor do they even have to resemble each other. Frequently, they do not. --EncycloPetey (talk) 05:03, 12 February 2015 (UTC)
  • The seventh item in Hrishikes' list is the question I wanted to ask, but was unsure of how to ask it.
Other details in this specific case include:
  1. This was one of the early works introducing general relativity. Spacetime as a single concept was not well-known then, so "space [and] time and gravitation" (the interpretation with the comma in place) would make sense.
  2. The title page plainly does not include the comma.
  3. The cover on our scan (from IA) has no text at all.
The second detail is the reason I concurred enough with EncycloPetey to make the move the first time.
The title is also set in all uppercase, so I could believe that the typographic convention of the time was to omit commas on title pages, even when the logical title includes them. EncycloPetey also pointed out that the filename used for proofreading and the title we use in mainspace need not actually be related, so while this question need not stall proofreading if we can all agree to just run with what we have, we will need an answer when it is time to publish this work in mainspace. Pathore (talk) 05:37, 12 February 2015 (UTC)
About the typographic convention of omitting commas on title page: this may be true as it is also omitted on the title page of Index:The Origin of the Family, Private Property and the State.djvu. Hrishikes (talk) 13:01, 8 March 2015 (UTC)

New to working over here on WS, but I've done a first run on a few pages of this. The OCR is quite good, so it's pretty quick to do, it would be nice if someone who knows more about the templates here would do the TOC (I'm a bit clueless about how to make it work). Revent (talk) 19:21, 18 February 2015 (UTC)

@Максим Пе: @EncycloPetey: @Pathore: @Koavf: Poking at people who expressed interest... I have done the first run-though of the entire work (including some huge chunks of TeX)... it's ready for validation (probably with some formatting cleanup) and transclusion. As I commented on IRC, it really needs a bit of annotation in some places, simply because of the historical context... as a specific example, he refers to 'spiral nebulae' that are now known to be other galaxies... at the time, it was thought that the Milky Way spanned the entire universe, and that other galaxies were simply distant nebulae in the Milky Way. Also, he talks about the number of planets, but Pluto was unknown at the time, and so what he wrote could be a bit confusing. Really, I think what needs 'annotated' would be a matter of a read-through by 'non-physics non-math' people, with pokes about what seems strange. His arguments were never 'wrong', even from a modern perspective, they just could use a bit of context in places. Revent (talk) 03:12, 11 March 2015 (UTC)
@Revent: When it comes to annotations, I'd recommend Wikibooks, especially if those annotations are long or involved or likely to change with new insights or if they could be enhanced by photos and video, etc. I'll be happy to poke around some validation. —Justin (koavf)TCM 03:23, 11 March 2015 (UTC)
@Revent: Also, it seems that Space Time and Gravitation from Author:Arthur Stanley Eddington is a redlink. Is this a deliberate choice? —Justin (koavf)TCM 03:25, 11 March 2015 (UTC)
@Koavf: I haven't attempted to transclude it yet, still new over here. I'll do so, but need to read the instructions a bit first. As far as annotations, I didn't mean anything extensive, just short notes where his 'terminology' would differ from what a modern reader would understand (like that 'spiral nebulae' refers to 'spiral galaxies', and that he didn't know they were external to the Milky Way... the book predates that discovery). Revent (talk) 03:49, 11 March 2015 (UTC)
@Koavf: I made an attempt at transcluding it, though it might well be broken... the entire work is one page, which I don't think is correct.... feel free to fix it and hand me a clue. :) Revent (talk) 04:04, 11 March 2015 (UTC)
I think I have it done correctly now (a page per each chapter, etc) at Space_Time_and_Gravitation. Feel free to fix it. Revent (talk) 07:43, 11 March 2015 (UTC)
We still have the issue of what the title actually is, specifically whether or not there should be a comma after "Space". See the above discussion, since this is the point where we can't put the decision off any longer. Pathore (talk) 21:02, 11 March 2015 (UTC)
@Максим Пе: @EncycloPetey: @Pathore: @Koavf: (again pinging people, sorry for the late response) Thoughout the work, the term is given as space-time, with a hyphen. I would suspect that the 'canonical' title would actually be "Space-Time and Gravitation", at least in the author's intent, though the actual title page and catalog entries seem to differ. Revent (talk) 05:46, 17 March 2015 (UTC)
So it does. Now we have three choices. Since the work uses "space-time" in the text, and the title page clearly has "space time", I suggest leaving it at "Space Time and Gravitation", possibly creating redirects from "Space, Time and Gravitation" and "Space-Time and Gravitation". Pathore (talk) 21:06, 17 March 2015 (UTC)
I also think that would be the best answer, with those redirects. Revent (talk) 22:30, 17 March 2015 (UTC)
I also agree that sounds like the best solution. --EncycloPetey (talk) 01:53, 18 March 2015 (UTC)

Is this a table?[edit]

I have no idea exactly how to format the second through fifth lines of Page:A dictionary of the Book of Mormon.pdf/106 because I can't be sure whether it is more or less intended to be in tabular form or not. Any ideas? John Carter (talk) 16:26, 9 February 2015 (UTC)

No, it's definitely not a table, and I believe that my edit to that page using {{ditto}} is probably all that is needed for this, even though it doesn't exactly line up. —Beleg Tâl (talk) 16:55, 9 February 2015 (UTC)
Good enough. Thank you. John Carter (talk) 17:06, 9 February 2015 (UTC)
Only a table would get the quotations aligned.— Ineuw talk 17:14, 9 February 2015 (UTC)
Actually, @Ineuw:, the template {{ditto}} will do that as well. --EncycloPetey (talk) 19:46, 9 February 2015 (UTC)
Added {{Ditto}} to Category:Special effects templates, if anyone feels this decission is counter-intuitive to the discovery of valuable templates please revert my edit. Actually on further investigation it seems {{Ditto}} doesn't line up as perfect as it could be.--Rochefoucauld (talk) 01:16, 10 February 2015 (UTC)
@EncycloPetey: There are many templates I am ignorant of and {{ditto}} is one of them . . . all I care about is that when the right tool is found I use it. It's mind boggling the number of ways one can format text in WS and if I would be studying and memorizing templates, that's what I would still be doing and I wouldn't have gotten any proofreading done.— Ineuw talk 00:11, 11 February 2015 (UTC)

Index:The_New_Latin_Primer_(Postgate).djvu[edit]

Not sure how to do the breve character on the pages still marked as problematic otherwise this one is proofread, but in need of validation.ShakespeareFan00 (talk) 18:28, 15 February 2015 (UTC)

Ancient large DjVu files no longer register as valid on Commons[edit]

I suspect the reason the following 3 files all-of-the-sudden stopped rendering properly (generating well over a thousand orphaned Page: namespace pages as a result) is because they used DjVu Digital or some other non-DjVu Libre [core] program to patch this or that prior to the latest/last uploaded version of the DjVu file.

Note the 0 × 0 file-size dimensions being reported by each; I've only seen this happen when some other DjVu-ish, non-Libre program has manipulated the file in some way (corrupted might be a more appropriate term imho).

Even if I could manage to download these, I doubt I could re-upload over them with "fixed" versions as Commons does not seem to like me in particular much when it comes to files over ~50M (or takes longer than an hour to upload).

Anyone with DjVu Libre familiarity willing to take these on? -- George Orwell III (talk) 09:46, 20 January 2015 (UTC)

p.s. - There is a newer version of EB1911, Volume 25 available on Internet Archive IN HERE. A straight replacement that is only ~2 years old vs. the existing one that is already ~6 years old -- even with bulk moves required -- is still better than fixing up 'old crap' imho. -- George Orwell III (talk) 03:59, 8 February 2015 (UTC)
N.B. c:File:EB1911 - Volume 25.djvu seems to be fine at present... did user:DutchTreat's upload of 7th February fix the issue; or is this in fact the wrong reference? I am asking because as I write both c:File:EB1911 - Volume 26.djvu & c:File:EB1911 - Volume 27.djvu are showing signs of distress. (I hope this isn't a "fix one get two bad replacements deal"?) 121.216.198.126 05:05, 12 February 2015 (UTC)
Volume 25 does indeed render a thumbnail now... but no auto text-layer dump (at least not for me). I know its there because I can OCR button it and that exactly matches what I get after I downloaded and extracted the layer on my hard-drive. Can you get a text layer upon entering/creating a new Page:? -- George Orwell III (talk) 05:14, 12 February 2015 (UTC)

EB1911 Volume 25 Scan File Corruption[edit]

I am having trouble with pages transcluded from scans of Volume 25 of EB1911: (transcription project). Is there corruption in File:EB1911 - Volume 25.djvu? Thanks for the help! - DutchTreat (talk) 11:42, 6 February 2015 (UTC)

File looks corrupted ... EB1911_-_Volume_25.djvu ‎(0 × 0 pixels, file size: 94.12 MB, MIME type: image/vnd.djvu). Probably needs someone to do a fresh load of a djvu file. — billinghurst sDrewth 06:28, 7 February 2015 (UTC)
I touched on this phenomenon up in this post. Vol. 26, Vol. 27, and EB1922 Vol. 31 all fall into the same group with the same issue as the other .djvu source file previously listed w/ Vol. 25 above btw. -- George Orwell III (talk) 20:23, 7 February 2015 (UTC)
... and don't ask me why the text layer isn't showing for the replacement vol. 25 -- its there when I Dl it to my hd. It probably needs a full [cache &] file refresh on Commons but THIS probably(?) prevented that from happening "normally". -- George Orwell III (talk) 03:36, 8 February 2015 (UTC)

Source file failures getting worse and worse every week[edit]

Just the latest Orphaned file list thanks to the source file registering as "flawed" all of the sudden. -- George Orwell III (talk) 21:29, 11 February 2015 (UTC)

Ugh. Browsing the Statutes at Large category on Commons makes me think this issue is confined to a (relatively) small number of volumes (for now…?). I’ll see whether I can re-download the original source PDFs from FDSys and re-convert the affected files. Tarmstro99 22:41, 11 February 2015 (UTC)
Don't go nuts just yet - its not just SaL volumes but most any series where a volume runs into the thousands page-wise. Originally I thought it was because the source file was too close to 100Mb in size but that seems not be the rule of thumb anymore.

This happened once before and the "memory was increased" or something to correct for it if I remember right. I think @Phe, Tpt: should be consulted before we get too far into "replacing" volumes at any rate. -- George Orwell III (talk) 22:52, 11 February 2015 (UTC)

I just replaced Volume 118 of the Statutes at Large with a new version that was converted from PDF using pdf2djvu instead of djvudigital, which was used to create the (broken) version of the file. The result was a slightly larger file, but it still does not work. This suggests to me that the problem may not lie with the conversion program but somewhere else instead. I’ll refrain from any further tampering until a better understanding of the actual technical problem comes to light. Tarmstro99 02:13, 12 February 2015 (UTC)
I tried to stop you @Tarmstro99: )

Anyway, the true test would be a.) set yourself the bot bit in your user rights; then b.) upload the file locally to en.wikisource rather than continued over-writes on Commons using a really different name to see if the same thing happens to insure the current state of the DjVu source file is not an effect of old garbage file caching or something isolated to Commons.

Depending on how that turns out; we still might need to manually test the text layer to see if that is behind it somehow instead. -- George Orwell III (talk) 03:13, 12 February 2015 (UTC)

┌───────────────────┘

I think I found the "crux" of the issue causing this. See the tracked task now linked to the right and somebody tell me if we're in the ballpark re: metadata size. -- George Orwell III (talk) 06:31, 12 February 2015 (UTC)

Close, but not quite. The URL given in that report returns an error, citing a limit of 12,582,912 bytes on results, while a similar URL for File:United States Statutes at Large Volume 118.djvu returns a 12,127,802 byte JSON response.Warning: ~12MB at that link. The returned XML has an incorrect DOCTYPE, but so does the returned XML for File:Kopal-Kundala.djvu, which has no problems and is our current Proofread of the Month. Pathore (talk) 01:50, 13 February 2015 (UTC)
But that's the 5th uploaded version of vol. 118 & not the necessarily the version that caused the log jam now replicated by [what I believe is] shite caching on commons. I forget which one does go over the limit (volume 123 or volume 120).

Why can't a 3rd party just upload a file in question locally to see if the same thing happens? I'll delete it right after either way. -- George Orwell III (talk) 20:54, 13 February 2015 (UTC)

Copying locally c:File:EB1911 - Volume 25.djvu — ( Index: here ), should be here soon--Mpaa (talk) 21:16, 13 February 2015 (UTC)
Not a big difference ... or were you expecting another file (I took the latest in Commons)?--Mpaa (talk) 21:57, 13 February 2015 (UTC)
Text layer looks OK:
user@pc:~/Downloads$ cp EB1911_-_Volume_25.djvu duff.djvu
user@pc:~/Downloads$ djvused duff.djvu -e 'output-txt' > duff.txt -u
user@pc:~/Downloads$ djvused duff.djvu -f duff.txt -u -s--Mpaa (talk) 23:34, 13 February 2015 (UTC)
I can't understand why the replacement EB1911_Vol._25 DjVu source file doesn't produce a text-layer dump in the Page: name space either.

Q:@Mpaa: you're the "PY guy", what did this have to do with the replacement file or replacement process exactly? -- George Orwell III (talk) 00:17, 14 February 2015 (UTC)

Wild guessing. My understanding is that some of Phetools are running on tools.wmflabs.org and check files as they come. E.g. hocr.py is run a few hours later than a new revision has been uploaded. I guess is that depending on check results, some further scripts might be run on a page basis. What they do I did not dig too much into. The code should be here: https://github.com/phil-el/phetools. Best is to ask directly @Phe:.--Mpaa (talk) 11:27, 14 February 2015 (UTC)
Could it be that if djvu is not considered to have a good quality by phetools, a new OCR needs to be done on demand via OCR gadget? I say this as MediaWiki:Gadget-ocr.js gadets points to "var request_url = '//tools.wmflabs.org/phetools/hocr_cgi.py?cmd=hocr&book=' ....", which is very similar to the above.--Mpaa (talk) 20:32, 16 February 2015 (UTC)
OK, a single page extracted from the djvu file works, so it has probably something to do with the total number of pages? See Index:EB1111 25 p45.djvu.--Mpaa (talk) 20:45, 16 February 2015 (UTC)
Volumes 120 and 123 are both broken and both show only a single version of the file on Commons. The analogous API query succeeds on volume 120, returning 12,892,030 bytes, but fails with an errorHTML JSON on volume 123, citing the same response size limit of 12,582,912 bytes. Obviously the limit doesn't work quite right. I'm waiting for the download from Commons to finish, then I'll do the tests on the text layer, although I don't see how a bad text layer could prevent the images from being recognized. Pathore (talk) 23:13, 13 February 2015 (UTC)
Its not going to be text-layer related. Even locally --> File:Vol118.djvu. -- George Orwell III (talk) 23:17, 13 February 2015 (UTC)
You had mentioned the possibility earlier, so I decided to try it. All I wasted is a few minutes of CPU time and I can now say that extracting, removing, and re-merging the text layer on volume 120 not only succeeds, but produces an output .djvu with the same SHA1 hash as the file I started with. Pathore (talk) 23:32, 13 February 2015 (UTC)
Confirmed (for the sake of completeness). While not "perfect", the text-layer extracts and [re]inserts without catastrophic failure (as was the case in this unrelated issue) for me too. Same results with my last two uploads of Volume 118 fwiw. Since volume 118 has 3496 pages comprising it; imho, I think we should stick to just poking volume 118 and volume 118 alone from now on. -- George Orwell III (talk) 00:17, 14 February 2015 (UTC)
I used volume 120 because there was only one revision of that file on Commons, rather than a bunch of them as for volume 118. (I couldn't figure out how to write an API request for metadata on an old revision.) I agree that we should focus on getting volume 118 to work, though.
For the wider problem, however, c:File:EB1911 - Volume 25.djvu mentioned above is particularly interesting: we have both a good and a broken revision before the problem was solved by uploading a newer IA scan. The broken revision was uploaded by Phe, who might be able to shed some light on the processing steps that produced it. I've compared djvudump output for the good and broken files and found some notable differences: (1) the second file has a text layer (of course), (2) while the image sizes match, the second file is marked as 300 dpi, while the first was 200 dpi, (3) the JB2 bilevel data blocks are different sizes, implying that the images were recompressed, and (4) the internal file names are longer in the broken version. Item (4) is less likely, unless Commons doesn't like DjVu files with pages named "page_####.djvu" (digits represented as "#") for some insane reason, because the new IA scan has even longer internal file names. The DjVuLibre tools here have no problem with any of the files, so I suspect something is wrong with the handling of .dvju files in MediaWiki. Pathore (talk) 03:52, 14 February 2015 (UTC)
I really don't get what is going on here; the same file as File:Vol118.djvu but with the text-layer completely removed (File:2vol118.djvu) seems to process and render just fine. Could the total number of pages with "hidden-text" exceed some threshold for the first test file rather than the total size of the "hidden-text" overall be somehow at fault here? -- George Orwell III (talk) 04:42, 14 February 2015 (UTC)
A sampling of files on Commons in c:Category:Statutes at Large turned up some good files with page names that didn't match the pattern "p*.djvu" (and none that did), so I tried rebuilding the bundle with different page names. Thankfully, the problem isn't quite that insane--it's still broken.
Too many pages with text is plausible: the three revisions of c:File:EB1911 - Volume 25.djvu have, in chronological order, 0, 1,090, and 1,085 pages with hidden text layers. The one with 1,090 text chunks is broken. The one with 1,085 text chunks works. There are maps on DjVu pages 25, 483, 519, 526, and 551 in that volume. The new IA scan doesn't have OCR data for those pages.
To extract a list of pages without OCR data from a bundled DjVu file: djvudump file.djvu | awk '/FORM:DJVU/ { if (page ~ /./) { print page } page = $3 } /TXTz/ { page = "" }'. A list of internal page names is printed, each in curly brackets on its own line.
-- Pathore (talk) 05:30, 14 February 2015 (UTC)
Further digging: most of the files in c:Category:Statutes at Large with more than 1085 pages do not have text in their metadata at all. The notable exceptions that I've found are 47 Part 2, 48 Part 1, 49 Part 2, 124, and 125. I got this by manually cross-referencing the results of this API query with API queries for metadata on the individual files that had more than 1,085 pages. I ignored the broken files when gathering this list. It looks like the issue is some combination of total text and number of pages. The working revision of c:File:EB1911 - Volume 25.djvu actually produces more output from djvutxt than the broken one. I'll look closer at those other volumes later; maybe they don't actually have as many text chunks as the API responses imply. Pathore (talk) 07:01, 14 February 2015 (UTC)

Error: Numeric value expected[edit]

I do not understand why I get this error on: Index:EB1911_-_Volume_26.djvu and Index:EB1911_-_Volume_27.djvu. Any clue?--Mpaa (talk) 20:22, 16 February 2015 (UTC)

You get that on the Index: page of just about any source file that is 'flawed' or 'failing'. Its a symptom of whatever is behind this latest 'failure to render' issue & not part of the cause itself.

This much I've been able to piece together:

  • DjVu files, specifically, store informative data components like it's own metadata, the text-layer & related mapping, annotations like URLs (rare) and the document's outline (or structure if you prefer) in one of two ways or some combination of two ways -- either as shared additional .iff files found within the bundle of indirect .djvu files that makes up the typical DjVu source-file for us or is applied/attached to every single indirect .djvu file found within a single DjVu source file. (to recap; each scanned page converted to a single djvu is an indirect .djvu file. When every scanned page have each been converted to indirect djvus and then complied into one, that is then defined as a bundled .djvu file).
  • When a .DjVu is uploaded and hosted as a source-file, all that additional, internal "informative data" is processed as / forced to the metadata part of the File: namespace where one typically finds info like camera-type, various date, dpi, colors, and similar for true image files like .PNGs, .JPGs and .TIFFs for example. In other words, since the .DjVu file specification can be considered slap-dash at best, instead of being thought first and foremost as a document file (mainly text; like a .DOC or .PDF) that is comprised of images of scanned "pages" -- some of which may actually be scans of illustrations that should be considered images -- its (erroneously) thought of MIME-wise as an image file. This is why developers "dump" what should be document metadata as image metadata. That box at the bottom of a File: page where image metadata goes has some "limit" apparently.
  • Its this questionable manner in the "hosting" of .djvu files with rather large or complex internal "informative data (or djvu metadata)" that seems to be catching up with us.
I haven't found the trigger point -- when exceeded -- that causes this "failure" we are all seeing. My last uploaded effort still hasn't crossed that threshold and its nowhere near the 4,000 [indirect djvu] pages total never mind only 2,000 of which have a text-layer (the additional informative data from before). -- George Orwell III (talk) 21:05, 16 February 2015 (UTC)
OK, thanks. I checked Index:EB1911_-_Volume_25.djvu and I was able to get the whole metadata via API and at first glance looks OK as XML file. I wonder if it more size than number of pages that matters (metadata is 11 Mb and it has "only" 1064 pages). If you need something, let me know, for me it is quite easy to retrieve info via API.--Mpaa (talk) 22:00, 16 February 2015 (UTC)
Whatever it is, the newer scans from the Internet Archive don't seem to trigger it. Volume 25 has been replaced with one of those, but volume 26 has not. You could try simply uploading the rest of the 2012 set of scans from IA to Commons if you just want the 1911 Encyclopaedia Britannica to work. Pathore (talk) 22:26, 16 February 2015 (UTC)
True; while the single trial of the replacement of EB1911 Volume 25 does seem to process "all the way" (e.g. has a file size, the height and width dimensions plus the page count), it still does not "automatically dump" what appears to be a perfectly good text-layer upon initial page creation in the Page: namespace. If this is true for other newer EB1911 volumes - I cannot say. Somebody with a faster connection should try replacing them just to see if Vol. 25 is an isolated case or not; we can always revert/delete/replace as needed afterwards.-- George Orwell III (talk) 22:53, 16 February 2015 (UTC)
EDIT: an exact copy of the current .php file dealing with DjVu "processing" is --> DjVuImage. Line 40 jumps out at me for starters. -- George Orwell III (talk) 23:33, 16 February 2015 (UTC)
That is a ~300MB memory limit. The djvutxt tool doesn't use anywhere near that much; 40M is adequate for processing all three versions of EB1911 volume 25 here.
Any chance we could get debugging logs from processing some of these files? There's a bunch of wfDebug calls in that code. Pathore (talk) 01:00, 17 February 2015 (UTC)
Beats me -- maybe Mpaa knows how?
No, sorry, no clue on that.--Mpaa (talk) 18:44, 17 February 2015 (UTC)
And Mpaa; can you extract that XML you mentioned before and put it up raw in sandbox or something for me please? TIA. -- George Orwell III (talk) 04:28, 17 February 2015 (UTC)
Let's try this link to Phab (let me know when you are done, so I can throw it away ...)--Mpaa (talk) 18:43, 17 February 2015 (UTC)
Thanks & sorry for the delay - I've got the XML now so you can do away with the Phab copy.

Right off the bat - I notice the internal indirect .djvu file name is not associated anywhere with its corresponding "object". And don't hold me to it but I do recall something about this "disassociation" occurring when some-caps/all-caps are used in the naming of the individual internal indirect .djvu files. I'll see if I can find it.

Moving on... is there anyway to script the renaming of these indirect .djvu files once extracted (I don't like position /0001 being ____0000.djvu in the EB1911 Vol. 25 file primarily). More important - is there an easy way to take those renamed indirect djvu files and make them back into a single bundled djvu without having to enter the name for each [renamed] djvu in the djvm command line? -- George Orwell III (talk) 06:13, 21 February 2015 (UTC)

1) about renaming, I think we can find a way. But I do not recognize your names, this is what I see:
djvm -l EB1911_-_Volume_25.djvu

61426 PAGE #1080 VOL25_SHWEBO-SUBLIMINAL_SELF_1079.djvu --> you woul like this to be SELF_1080.djvu?

2) after renaming, looping djvm -i[nsert] doc.djvu page.djvu [pagenum] should do the work. or?--Mpaa (talk) 19:54, 21 February 2015 (UTC)

┌─────────────────────────────┘
C:\Program Files (x86)\DjVuLibre>djvmcvt -i EB1911_-_Volume_25.djvu c:\tempwork v25indx.djvu

Fwiw.... the above will extract the bundled djvu into 1,090 indirect djvu files plus 1 additional "index file" (v25indx.djvu - should not be carried over in renaming btw). Once extracted, we should try a renaming scheme taking VOL25_SHWEBO-SUBLIMINAL_SELF_1079.djvu to something all lowercase & shorter like eb1911_vol25_1079.djvu for example. How to take all the renamed files and combine them again (I guess?) would be a djvm command with 1090 variables or 1090 djvm commands starting with the last renamed, indirect djvu and ending with the first. I'm still wary of any file using 0000 instead of 0001 for the first position however. -- George Orwell III (talk) 20:56, 21 February 2015 (UTC)

I've tried this already with a different file, and did not get a very useful result. The internal DjVu page names don't seem to be used for anything in MediaWiki, but I can do these sorts of rebuilds very easily. Uploading the file afterwards is annoyingly slow (limited bandwidth) but doable. I don't think that the page numbers are in any way tied to the internal file names. Also, note that the version of EB1911 vol. 25 where the internal files are numbered from 0000 is one of the files that works. Pathore (talk) 21:30, 21 February 2015 (UTC)
None of this makes any sense (going by what DjVulibre provides). Between inspecting the XML that Mpaa pulled earlier and what is currently found in DjVuImage, it seems that MediaWiki is doing twice the work that it should be (or needs to?) -- first by listing all the "objects" [the internal indirect djvu files] & their dimensions under <DjVuXML> followed by a per-page text dump under <DjVuTxt> -- the latter being the one with the internal file position -to- file name info when that info should be listed in the former. While its understandable to initially omit the text layer to insure the reporting of file size, dimensions and page count successfully executes first, at the same time there is absolutely no reason to leave out the position & page name in <DjVuXML> the way I see it. And the redundancy of compiling <DjVuTxt> afterwards is a "hack" at best; the proper way to limit the level of "detail" in the stored text layer to PAGE would be to edit the associated DjVuXML-s.dtd file to reflect that as the default instead of LINE or WORD. Or did I just confuse everybody? -- George Orwell III (talk) 22:01, 21 February 2015 (UTC)
Uploading of File:EB1911_-_Volume_25.djvu on its way. Why do you say that EB1911 vol. 25 works? It does not load text layer when I try to create a page.--Mpaa (talk) 22:19, 21 February 2015 (UTC)
Awaiting upload @Mpaa: . In the meantime an FYI - New version of DjVuLibre released 2/12/2015 MORE HERE. I don't know what, if anything, has been improved or changed. Please report if anybody finds something "new" -- George Orwell III (talk) 22:32, 21 February 2015 (UTC)
Volume 25 works in the sense that the page images are there. We also have problems with other files where even the pages are not recognized and the file is reported as being 0x0 pixels. File:EB1911_-_Volume_26.djvu currently has this latter problem. Pathore (talk) 00:20, 22 February 2015 (UTC)
Yeah but Volume 25 had the same issue(s) at the outset that are affecting vol. 26 & 27 now so there is little to gain by spitting hairs here. I do think we are at the point now were we should focus on just one or two test cases suffering from this phenomenon and bot import text from IA's dump file along with new Page: creation for the remaining Index:es in trouble. Thoughts? -- George Orwell III (talk) 02:56, 22 February 2015 (UTC)
In short, we have multiple problems. One bug is causing some text layers to not show up, while another (or the same bug in another context) is causing DjVu files to be rejected in their entirety. As for bot imports, I've been thinking a bit about writing a semi-automated tool that would use the DjVu text position annotations to separate text into headers, page bodies, footers, paragraphs, etc. and then automatically bulk-upload Page:s, complete with a special template for flagging words not in its dictionary as suspected OCR errors. Pathore (talk) 04:34, 22 February 2015 (UTC)
I believe that kind of "sectioning" is what DjVutoXML was intend for. Its this stupid circumventing of the creation of that DjVuLibre XML file for DjVu-txt or DjVu-dump that has irked me for years now. Here is an example of a proper XML output, complete with mapping for columns, regions, paragraphs, lines and words. -- George Orwell III (talk) 05:11, 22 February 2015 (UTC)
That's close, but I have also been thinking about automatic detection of pictures and diagrams, so that, after the upload, all pages are either "not proofread" if they contain only text, or "problematic" if they contain marks in areas where there is no text. I'm still turning the architecture overview over in my head on this. Pathore (talk) 05:20, 22 February 2015 (UTC)
@Pathore: - Its a matter of properly determining the x & y min/max coordinates for every element, No?

If we have the entire PAGE's dimensions and the number of COLUMN or COLUMNs set, I think all we'd need is a way to detect REGION(s) substantially differing coordinate wise from other REGION(s) with the same COLUMN parent & PAGE; such a shift in x &/or y min/max would probably be a good sign that an image or line drawing resides within (opposed to the norm containing portions of the hidden/embedded text-layer). -- George Orwell III (talk) 05:54, 22 February 2015 (UTC)

I was thinking more of starting with words (or even characters) and reconstructing the larger groups by matching up bounding boxes. The larger region information in DjVu isn't something I've ever heard of anyone actually relying on, so I would be somewhat reluctant to trust it. After building up this kind of "box model", determining whether a gap between text boxes is a figure or just whitespace becomes a simple matter of decoding the DjVu image and checking if the possible figure has characteristics similar to the margin boxes. Pathore (talk) 06:07, 22 February 2015 (UTC)
Sounds good to me too & you have a point regarding the use & support of the "higher" containing elements.... but characters and/or words would be just as problematic from my experience -- frequently punctuation is incorrectly detected and deemed WORD-like for example. Even stray specks or pen marks are picked up and assigned WORD-like status. LINE is the obvious choice here imho.

A LINE's slight indentation compared to other LINEs would be a decent marker for the equivalent of a Paragraph start just as a LINE less than the right-hand x-max (typically justified anyway) would be a good indicator of paragraph end [logically followed by a new paragraph start or if unusual offsets = could indicate your line drawing /image from before]. Colons, mdashes and the like would automatically collapse any errant whitespace/hair-space if LINE were used as the "lowest" element over WORD as well.

Either way, the current .phps dealing with DjVu's feel more like handicaps than helpers and that's probably where an overhaul is needed before addressing fanciful things like what we're discussing. -- George Orwell III (talk) 06:28, 22 February 2015 (UTC)

I'm planning to write this as a tool running locally on my computer for more-or-less that exact reason. This kind of processing simply isn't something for which PHP is well-adapted. Nor does the tool have to be perfect; that's why we have proofreading. But a Page: extractor won't be of much use if the server doesn't even recognize that the pages exist, so we do need to get the DjVu handling fixed. Pathore (talk) 07:41, 22 February 2015 (UTC)
yes, please. vols 1 & 25 seem to no longer have a text layer; vols 26 & 27 have "numeric" error. it would be nice to have these sorted by year-end, so we can have all the volumes "proofread ready". Slowking4Farmbrough's revenge 18:48, 23 February 2015 (UTC)

Incomplete author page[edit]

Can anyone please add ascertain the death-year for this author page and both birth- and death-years for this author page and add them? Thanks. Hrishikes (talk) 03:17, 17 February 2015 (UTC)

???I'm confused. What is preventing you from doing it exactly? -- George Orwell III (talk) 04:25, 17 February 2015 (UTC)
I'm confused at your confusion. I'm asking for help; which means I'm unable to do it myself; which again means I'm unable to get the relevant data. Hrishikes (talk) 04:43, 17 February 2015 (UTC)
Oh, I think I understand now. You don't actually know these dates and need help finding them out; not so much 'need help adding some dates to the author: namespace' (I edited your opening post to better reflect what I think you're asking help with). -- George Orwell III (talk) 04:53, 17 February 2015 (UTC)
Chowdhury died 1939, and I have transcribed the Times obituary to the talk page. No indicative year of birth, though I would have said prior to 1880s at a minimum. For Sen, I have looked previously without results. I don't have good sources for India, the National Archives aren't useful, and I find the National Library better to use, though not necessarily helpful. — billinghurst sDrewth 13:59, 17 February 2015 (UTC)
Thanks a lot; I don't know why this did not come up in my search. Hrishikes (talk) 14:47, 17 February 2015 (UTC)

The Pilgrims' Halt[edit]

I'm not sure how the poem The Pilgrims' Halt is formatted. --Lo Ximiendo (talk) 12:11, 17 February 2015 (UTC)

Had a try. Please check. Hrishikes (talk) 13:11, 17 February 2015 (UTC)
Tyvm, Hrishikes. :) --Lo Ximiendo (talk) 13:35, 17 February 2015 (UTC)

Index:The New Latin Primer (Postgate).djvu (pp 196, 198).[edit]

A fairly simple task if anyone is able to assist.

In the text for these pages is some typesetting for Prosody, but I was not happy with the rather hasty kludge I'd used, hence the ? in the page. Can anyone suggest what the appropriate symbols/notation this work is using? (And whether those symbols are supported in the version of Unicode Wikisource users are likely to be using.) ShakespeareFan00 (talk) 12:32, 17 February 2015 (UTC)

These are macron, breve etc. You may like to see the article w:Foot (prosody). Hrishikes (talk) 13:10, 17 February 2015 (UTC)
Do you mean things like o͝o, where the breve tops two successive letters? I'm not sure how that's done, but I do know that on Wiktionary we've had to contend with the double-O version I pasted in. Someone there might know more. --EncycloPetey (talk) 05:14, 23 February 2015 (UTC)
I think you will find that is a use of the UNICODE "combining double breve" (html entity &#x035D)? Something like o&#x35D;o ought to produce similar output to your example (i.e. o͝o).

Page:Adapting_and_Writing_Language_Lessons.pdf/401[edit]

In the dictation, some of the symbols have a vertical line, what is this symbol and how can it be entered? (I'd appreciate some assistance on this work, as I also found some IPA like symbols I can't match up easily.)

It would be nice to get this one finished. ShakespeareFan00 (talk) 00:50, 18 February 2015 (UTC)

It's a w:Glottal stop. Transcribe it with an apostrophe (not curly). Beeswaxcandle (talk) 07:23, 18 February 2015 (UTC)
Follow on - Page:Adapting and Writing Language Lessons.pdf/193 . This has what looks like a reversed c and an n with a tale, the latter I'm fairly sure about, the reversed c I'm not. ShakespeareFan00 (talk) 09:41, 18 February 2015 (UTC)
These are IPA symbols, and can be found at w:International Phonetic Alphabet. Hrishikes (talk) 10:19, 18 February 2015 (UTC)

Author's name change needed[edit]

I need a name change from "Charles L. Lewis" to Charles Lee Lewis. —Maury (talk) 02:58, 18 February 2015 (UTC)

Yes check.svg Done . I've left a redirect from the initial version. Beeswaxcandle (talk) 07:27, 18 February 2015 (UTC)
Maury, you have the possibility to move the page by yourself, see Move under More tab in upper right side.--Mpaa (talk) 20:54, 18 February 2015 (UTC)

US copyright for non US works, some questions[edit]

Hi to all!

Sorry for my not very well Engilsh, but maybe somebody can help me providing with information on US copyright laws:

There is one unclear for me point on applying US copyright and copyright terms to non-US literary works (more precisely — to Russian / Soviet works), concerning URAA's restoration of copyrights for foreign works. The point is:

Let's suppose that we examine copyrights of some literary work which falls under conditions:
1) the work was first published out of the US;
and 2) the work has never been published in the US (so it was not published in the US during 30-day period following the publication date, as well);
and 3) the work was still under copyright in its home country as of the URAA date (1/1/1996 in the case of Russia, and many other countries);
and finally 4) It was published in its source country without copyright notice, and the publication was in the period when such notice was required in the US to establish the copyright

The question is: did this work fall under US copyright because of copyright restoration according URAA? And if yes — what term is applied for US copyright protection? As I've understood, the missing copyright notice does not prevent the copyright restoration, and at any case, any restored work is granted copyright protection despite missing US copyright notice and/or US copyright renewal (even if such renewal was required in the case of any US work). Am I right? Or I am wrong, and missing copyright notice makes obtaining US copyright to fail, and so far — the work is PD in US now? (and we can freely use it in the Wikimedia projects?)

Also please help to figure out copyright issues for following two cases:

  1. Some Soviet author published his work at 1931 year, the work was published without copyright notice. Years passed, and the author died in the USSR in the autumn of 1941 year. According to the Russian Civil Code, this work is under Russian copyright until 2015 year (inclusive), and enters PD in the 2016 year. The work was under Russian copyright on the URAA date 1/1/1996. The work has never been published (and has never been registered copyright) in the US, and so it has never got any copyright renewal. The questions: a) is this work under copyright protection in the US? — and if yes — b) what is term of the protection, what length does it have?
  2. Some Soviet author died at 1931. Some work of this author was firstly published only in the 1944 year (PMA). According to the Russian Civil Code, this work was under Russian copyright until 2014 year (inclusive), and entered PD this (2015) year. The following details are the same as in the previous case: the work was under RF copyright on 1/1/1996; in has never been published and never been registered copyright in the US. The questions are the same: a) is this work under copyright protection in the US? — and if yes — b) what is term of the protection, what length does it have? --Nigmont (talk) 20:28, 18 February 2015 (UTC)
Seems like both will be 95 years after publication from your description. Reading material Cornell copyrightbillinghurst sDrewth 12:08, 19 February 2015 (UTC)
  • Thank you very much! --Nigmont (talk) 18:14, 19 February 2015 (UTC)
Just as a 'pedantic' note (that might be worth mentioning), the described works would have been in the public domain in the US 'pre-URAA' not due to a failure to comply with the formalities, but due to a 'lack of national eligibility' at the time of publication... the Soviet Union was not a party to any international copyright treaties until 1973 (when they joined the UCC), so purely 'Soviet' copyrights were not recognized under international law before that time. Revent (talk) 10:41, 20 February 2015 (UTC)
Sorry for too late reply.
Revent, it would be good if you were right (I might put to Wikisource many works if so), but I am afraid that you are wrong. I think that 'eligibility' of the country is determined as of current status of copyright treates established between the US and the country, but not as of date where the work was firstly published. So it does not matter whether the Soviet Union was eligible on the date of publishing, but it does really matter that the Russia (the legal successor of the Soviet Union) is definitely an eligible party for US as of current status.
Moreover, the URAA act clearly states that the lack of the national eligibility is one of the conditions for 'copyright restoration' of foreign works - see Title V, SEC. 514. RESTORED WORKS, paragraph "(h) DEFINITIONS.", subparagraph "(6) The term restored work means an original work of authorship that", and following:
(C) is in the public domain in the United States due to ...
(iii) lack of national eligibility;...
So it seems that the US copyright term for such works is 95 years, the same as for US works having copyright notice and renewal. --Nigmont (talk) 18:11, 17 March 2015 (UTC)
@Nigmont: I was not attempting to say that they are not possibly eligible for restoration under the URAA, what I was pointing out was that they actually would fall under the criteria you noted, "lack of national eligiblity", instead of the criteria "failure to comply with the formalities". For a work from a source nation that was not a 'treaty partner' at the time of publication, but later became an 'eligible country', the lack of compliance with the formalities is irrelevant.... a Soviet work from the 1930s or 1940s would not have have had it's copyright recognized in the United States, even if it was published with a copyright notice that met the requirements of various international treaties, because the Soviet Union was not a party to those treaties at the time. That the Soviet Union decades later joined the UCC would not have granted those works 'retroactive' recognition, only the much later passing of the URAA potentially did so. Revent (talk) 20:44, 17 March 2015 (UTC)
@Revent: sorry for my mistake: I misunderstood your first comment here and wrongly assumed that you meant that those works should not be restored copyright in the US under URAA definitions. Now I understand that you really told some other things, thanks for your clarification now.
But nonetheless IMO you are not quite right when you said in your second reply: only the much later passing of the URAA potentially did so. I think that those works should be considered as not potentially but definitely restored with URAA (except cases when a work is a pre-1923 publication). At least, IMO, this point should be believed regarding the possibility of exposing those works to Wikisource which (similarly to all other Wikimedia projects) is more prone to 'copyright-paranoia' than to 'copyright-boldness'. So IMO these works should be believed as not in PD and so they may not be published on Wikisource on PD terms. See also link Status of the "rule of the shorter term" in the US (I found this in the multilingual Wikisource). --Nigmont (talk) 16:49, 18 March 2015 (UTC)
Yes, my use of the word 'potentially' was quite deliberate, in order to avoid making a (possibly controversial) assertion that the URAA applied to a specific work without looking at the details of that work.. URAA arguments are unfortunately common and heated on Commons. I do agree that, particularly for a work to be on Wikisource, it should have a copyright clearance that 'positively' establishes that it is PD (and not restored), rather that assuming that things are PD unless proven otherwise. The URAA should be interpreted to apply to 'classes' of works based on their age and origin, unless demonstrated otherwise, instead of assumed to not apply unless someone can argue that it does. Revent (talk) 10:13, 19 March 2015 (UTC)

What does this mean?[edit]

I was reading an entry for a John Hawkins M.D. and noticed that it was followed by (fl. 1635). What does (fl.1635) mean?

here you go, see Floruit --Rochefoucauld (talk) 03:50, 20 February 2015 (UTC)
It means flourished and refers to period when the person was active. But, huh, recently I thought that flourish means 'turn into flour', or turn into dust, so this abbreviation should refer to the point of time when the person vanished. LOLsmiley --Nonexyst (talk) 23:44, 22 February 2015 (UTC)

Index progress bar[edit]

Is there a way (e.g. a template) to show the progress bar of one particular index as it’s being displayed on the Special:IndexPages page? Nonexyst (talk) 14:36, 22 February 2015 (UTC)

See {{PageStatus}} or {{Progress}}. Not with automatic update as far as I know.--Mpaa (talk) 17:02, 22 February 2015 (UTC)
Thanks. Well, then, is there an expression to get the number of pages with a specific prefix in a specific category and an expression to get the number of pages in a specific PDF/djvu file? Nonexyst (talk) 18:05, 22 February 2015 (UTC)

Index:Egyptian self-taught (Arabic) (1914).djvu (p6,7)[edit]

Anyone here familiar with Arabic script? I've done most of this work, but would appreciate someone that knows the arabic scripts assisting in the construction in the tables on these pages.ShakespeareFan00 (talk) 22:15, 23 February 2015 (UTC)

You can apply to people listed in Category:User ar. Hrishikes (talk) 12:18, 28 February 2015 (UTC)

Possible Cite.php bug encountered with Table 4-12 in Index:NIOSH Hazard review of Carbonless Copy Paper.pdf[edit]

I've transcluded the full table together in the Sandbox, but there I'm getting "Cite error: Invalid <ref> tag; no text was provided for refs named table-4-12-1dd". The reference in question (along with the table's other footnotes) is defined in a <references> block on Page:NIOSH Hazard review of Carbonless Copy Paper.pdf/117, which is also transcluded into the footer on PDF pages 97 and 98. The error also appears on those pages, along with a bunch of errors about unused references, since each page of the table only uses some of the footnotes.

The reason I think I've hit a bug in Cite.php is that the "missing" reference is also used in two of the other footnotes, and renders without issue on the table's footnote page. Pathore (talk) 02:35, 24 February 2015 (UTC)

Thanks to George Orwell III's efforts, the table looks a lot better now (the footnotes are now numbered left-to-right across the top in the transcluded table). Since I didn't like hardwiring "reference" links, I kept looking for a better solution and this edit seems to have worked around the underlying problem. Special:ExpandTemplates showed that <includeonly><references></includeonly> ... <includeonly></references></includeonly> was being transcluded as <references></includeonly> ... <includeonly></references>; that is, transclusion is not strictly text substitution and incorrect results are silently produced if partial transclusion tags do not properly nest with other tags. Pathore (talk) 23:46, 24 February 2015 (UTC)

'Wikisource does not have a page with this exact name'? Oh?[edit]

My user name is "MisterCat" and I've been registered here for a few years now, or so I thought. For reasons best known to persons other than myself, my user name now appears in red and the admonition "Wikisource does not have a page with this exact name" presents itself when I click on the newly-reddened "MisterCat" hyperlink. My profile, on the other hand, thinks I'm duly registered and even offered to consolidate all "MisterCat" appearances site-wide; and so I followed up on that offer. Supposedly, that task was accomplished; but I'm still the red "MisterCat" here, the guy who allegedly hasn't a profile page. (But I do have one! Honest!) What can I do to stop being red and sending seekers, including myself, on a snipe hunt? Your courteous and helpful advice, dear readers, will be greatly appreciated. Thank you! -- MisterCat (talk) 20:07, 26 February 2015 (UTC)

You need to create the page User:MisterCat. that's where your user page for the English Wikisource will be located. Until that page exists here, the link from your name will be red. It has nothing to do with your account; the red link indicates the absence of a particular page. --EncycloPetey (talk) 20:29, 26 February 2015 (UTC)
Easily fixed. (and I am not even logged in!) 124.184.180.188 22:26, 26 February 2015 (UTC)
However, it is best practice to edit your user page while logged in, so that future admins know for certain that your page was not "vandalised" by an anonymous user. --EncycloPetey (talk) 22:30, 26 February 2015 (UTC)

Can an image taken from a historical US official book be uploaded?[edit]

The image I want to uploaded is from a book called "Foreign relations of the United States, Conference at Cairo and Tehran, 1943". It was one of the Foreign Relations Series, which are collections of historical documents of foreign relations from United States Department of State and are published by United States Government Printing Office. The series are now hosted on University of Wisconsin and can be freely read online. The one I'm referring to is this one, and this is its publishing (1961) information.

The page I want to upload is page 640. Can it be uploaded? If so, can I cut the page horizontally in half with Photoshop and only upload the bottom half? --Matt Smith (talk) 09:43, 28 February 2015 (UTC)

I don't see any "image" on P. 640. You do realize a scan of a printed page containing text might be considered an image file, its the content (the text) that is the "focus" as far as copyright protection(s) are concerned.

Since the original is a product of the U.S. [Federal] Government - its excluded from copyright protections and thus, is in the Public Domain. Its ok to do with as you wish in short. -- George Orwell III (talk) 09:54, 28 February 2015 (UTC)

Yeah, you are right. Sorry for not being clear. What I actually want to upload is "the image scan of p. 640".
Thank you for helping me fathom the copyright question. To be honest, I just realized that I should have asked this question on Wikimedia or Wikipedia because the place I want to use the image at is on a Wikipedia page. But I still thank you for helping me at here. Cheers.--Matt Smith (talk) 10:23, 28 February 2015 (UTC)

remove boilerplate before adding new text[edit]

I would like to add the original 1847 edition of "History of the press of western new york" by Frederick Follett to en.wikisource.org. IA has the 1920 and 1973 reprints but not the original 1847 edition - books.google.com has the 1847 edition about 80 pages. I would appreciate help from someone who can remove the google boilerplate and upload to IA to create the djvu file. There was a discussion of this topic on https://en.wikisource.org/wiki/Help:Internet_Archive/Requested_uploads which seems to be an old thread that no one reads. Robin2014 (talk) 16:19, 28 February 2015 (UTC)

Yes check.svg Done https://archive.org/details/historyofpressin00foll_1 Hrishikes (talk) 17:58, 28 February 2015 (UTC)

How do I cancel specific electronic-mail notifications?[edit]

So ... how do I cancel specific electronic-mail notifications? I recall seeing a box (which I'd ticked) asking if I wanted to be notified of changes, but I can't find that box anymore and I've been unable to find a box asking if I want to NOT be notified. Your help in this matter will be greatly appreciated. Thank you! -- MisterCat (talk) 18:58, 1 March 2015 (UTC)

https://en.wikisource.org/wiki/Special:Preferences#mw-prefsection-echo should be the User: preference page you're looking for... if not -- try https://en.wikisource.org/wiki/Special:Preferences#mw-prefsection-personal as well. -- George Orwell III (talk) 19:09, 1 March 2015 (UTC)
Thank you very much, George Orwell III, for your help! It looks like one either accepts ALL e-mails or cancels them all, so I've cancelled them. Alas! -- MisterCat (talk) 20:25, 1 March 2015 (UTC)
If all else fails, disable JavaScript, then go to Special:Preferences. (The fancy tabs rely on JavaScript; without it, everything is displayed on one long page.) There should be a "Notify me about these events" box near the bottom, with two columns of checkboxes, one labeled "Web" and the other labeled "Email".
The single checkbox you seem to have found under "Email options" is to allow or disallow other users to use Special:EmailUser/MisterCat. I don't think it affects notifications. Pathore (talk) 02:03, 3 March 2015 (UTC)

Transcribing a tune[edit]

Oregon, My Oregon.jpg

Hi, I'd like to transcribe the state song of the U.S. state of Oregon. Can somebody familiar with such things help me get started? -Pete (talk) 22:53, 5 March 2015 (UTC)

I've created an index page for it: Index:Oregon, My Oregon.jpg. For the transcription, a good starting point is Help:Sheet music and Help:LilyPond. See Faith of Our Fathers for an example of a similar transcription. —Beleg Tâl (talk) 23:15, 5 March 2015 (UTC)
Marvelous, thank you! I'll take a look. -Pete (talk) 23:17, 5 March 2015 (UTC)

Investigation of the Ferguson Police Department (2015)[edit]

Hello, I just uploaded this text to Wikisource, but it's quite big and I'd like some help with it. If anybody can help in proofreading it and transcribing it, that'd be much appreciated. Thanks! Also, there's a PDF version of this uploaded here as well, but since there's a superior DJVU version, could somebody delete the PDF version? Thanks. Illegitimate Barrister (talk) 18:54, 6 March 2015 (UTC)

Good text, thanks for starting this. Working on it a bit now. -Pete (talk) 20:21, 6 March 2015 (UTC)
Here's another accompanying work. Illegitimate Barrister (talk) 11:03, 7 March 2015 (UTC)

I've (a it tendentiously, I guess) run though all 100+ pages of this. I'm currently going back through to make the formatting of the section headers (which was a bit odd) uniform thoughout the work, but it's ready for verification of the transcription... the OCR was quite good in most places, other than fixing apostrophes, quotation marks, and various types of hyphens. Revent (talk) 05:38, 17 March 2015 (UTC)

Just as a note, I'm intending to also do the Michael Brown one next, mainly because I'm interested in reading it, and proofing it at the same time isn't much of a burden. :) Revent (talk) 06:02, 17 March 2015 (UTC)

Same reference for three refs?[edit]

On this page there are three reference marks (*) pointing to the same reference. How can I resolve this?Ineuw (talk) 03:34, 7 March 2015 (UTC)

Name the reference and then re-use it. See Page:Ante-Nicene Fathers volume 1.djvu/162 for an example. Beeswaxcandle (talk) 03:46, 7 March 2015 (UTC)
The Ante-Nicene Fathers case is rather different to this one, in that that contains two references to the same text fragment. However if you read the context of the PSM page there is in fact one reference only which in turn contains another fully resolved sub-reference. I believe the "bar" separator at the bottom of the page ought to be incorporated into the text of the reference expansion. The clue to all of this is the lead-in discussion which reads "aggregate membership of nearly nine thousand". Clearly this refers to the table as it is a statement of fact and not in fact an estimation as implied if the (nested) reference is treated as primary.

Obviously some interpretations differ from mine. 121.216.134.246 07:40, 7 March 2015 (UTC)

I just "fixed" it -- hopefully if you look at the diff you can see the pattern. Note that you don't need quotes if there are no spaces in the ref's name, but you do need quotes if there are. (So, e.g., <ref name="any name" />) Also, note that the trailing slash is used only after the reference has been defined (i.e., the 2nd and 3rd instances, but not the 1st.) -Pete (talk) 03:52, 7 March 2015 (UTC)
Salute you both. Thank you.Ineuw (talk) 04:03, 7 March 2015 (UTC)
Oops. Crashed into above edit. Somebody please sort out the wreckage. Oh the humanity! 121.216.134.246 04:21, 7 March 2015 (UTC)

I think it is pretty clear it is a nested ref i.e. the footnote has a footnote. I have nested them; submitted for consideration; revert me if you disagree. Hesperian 11:08, 7 March 2015 (UTC)

This poor page has been hung, quertered, suffocated, garroted, shot, roasted on fire, and finally drowned. Why can't the community leave this poor page alone as it was recommended originally — with named anchors???Ineuw (talk) 03:58, 8 March 2015 (UTC)
Well my edits neatly cancelled themselves out, so if you want to clean up and remove them no skin off my nose. AuFCL (talk) 04:12, 8 March 2015 (UTC)

why was Military%E2%80%93Industrial_Complex_Speech deleted[edit]

why was http://en.wikisource.org/wiki/Military%E2%80%93Industrial_Complex_Speech deleted?—Preceding unsigned comment added by 73.201.72.72 (talk) 13:22, 7 March 2015

As far as I can tell, that page has never existed. I think the page for which you're looking is one of the links from Eisenhower's farewell address. - AdamBMorgan (talk) 15:17, 7 March 2015 (UTC)

Why is Creative Commons Attribution-ShareAlike 4.0 International License missing in the Licensing of the Upload-Wizard?[edit]

There's only ShareAlike 3 & 2.5 -> why? Couldn't find any information on that. I was able to select the 4.0 License in hindsight, but it's missing from the Licensing ListBox. To me this seems to be an error. --Fixuture (talk) 14:26, 7 March 2015 (UTC)

upload wizard is relatively unsupported. you would think they would jump to build on the good work of user:mindspillage, but alas not. i’ve been going to c:Commons:VicuñaUploader. Slowking4Farmbrough's revenge 21:07, 9 March 2015 (UTC)
Looking at the config item (wgUploadWizardConfig) in http://noc.wikimedia.org/conf/highlight.php?file=CommonSettings.php cc-by-sa-4.0 is allowed. If it is not coming through as a choice, then it should be a phabricator: request. — billinghurst sDrewth 12:08, 10 March 2015 (UTC)
I think I added the 4.0 series to the list so that it is a working option for all [local] uploaders now. Please verify. -- George Orwell III (talk) 00:06, 11 March 2015 (UTC)

Moving an image from Wikipedia to Commons[edit]

If there is someone with the technical skill and permissions to do a proper transfer of w:File:LYSISTRATAstaging.jpg to Commons at File:LYSISTRATAstaging.jpg, I would appreciate it. I can only transfer files the old fashioned way: download locally, then re-upload, which is a complicated process that doesn't preserve the edit history from the original location. --EncycloPetey (talk) 21:53, 8 March 2015 (UTC)

Yes check.svg Done , using tool: w:User:This, that and the other/For the Common Good. -- Cirt (talk) 22:16, 8 March 2015 (UTC)
Thanks. The documentation for that tool doesn't encourage me, as it's described as "most likely" working on my MacOS. --EncycloPetey (talk) 23:09, 8 March 2015 (UTC)
Ah, no worries, EncycloPetey (talkcontribs), good luck with it! :) -- Cirt (talk) 23:22, 8 March 2015 (UTC)
Very much so. You can see at Commons all the pages where I've already linked it, both here in the Portal:Ancient Greek drama and elsewhere across the Wikipedias. unsigned comment by EncycloPetey (talk) .
Excellent. I wasn't aware of this tool, will check it out. For these purposes I use Commons Helper, which lives on the Tool Labs site (and thus doesn't require you to run anything locally); I don't think it requires any advanced permission on either site (if you're not an admin, it gives you a nice easy button to request speedy deletion). There's also a nice script (which you can install at Meta, making it available on all wikis) that will give you a lefthand nav bar link "move to commons". See relevant links at the bottom of my blog post: [1] -Pete (talk) 22:25, 9 March 2015 (UTC)
ps @EncycloPetey: I have now deleted the local description page on enwp. -Pete (talk) 22:28, 9 March 2015 (UTC)
Great tools, Peteforsyth (talkcontribs), thanks for the helpful suggestions! :) -- Cirt (talk) 22:43, 9 March 2015 (UTC)
No problem. It's amazing how many useful tools go unnoticed in this wacky wiki world -- it's always hard to find the right one for the task. FYI, @Magnus Manske: also created a tool that permits the opposite -- moving a file from Commons to a local wiki -- useful in (relatively rare, but nevertheless significant) cases where a file is not permissible on Commons, but is permissible by a local wiki's policy. (Would have been useful, for instance, in the recent case you and I looked at, if the prevailing sentiment at Commons had been that the file should be deleted there.) I have used it, but I'm having trouble locating it or remembering its name right now. -Pete (talk) 01:26, 10 March 2015 (UTC)
Note: With "For the Common Good" if you need a configuration file for transferring from enWikisource to Commons, I have a version for download at User:Billinghurst/tools/FtCG configuration file for enWS.wiki. If moving multiple files, or needing to rewrite descriptions/templates on the fly, it is the preferred tool; if you are just looking to move one, then toollabs:CommonsHelper and as is an OAuth tool. — billinghurst sDrewth 11:05, 10 March 2015 (UTC)
Thanks @Billinghurst: - an ability to move multiple files and deal with changes to the description page is very attractive; I will try using it for the Centennial History of Oregon images discussed in the blog post/screencast above. Thanks for explaining! -Pete (talk) 15:18, 10 March 2015 (UTC)

I have made excellent PDFs of Public Domain books.[edit]

Is there a way I can donate my PDFs to Wikipedia so that anyone who wants to can download them?

The Wikimedia foundation's media are hosted at the Wikimedia Commons. You may be able to upload them there. Abyssal (talk) 20:21, 11 March 2015 (UTC)
It might work best if you first upload them to the Internet Archive and then, after IA processes them, upload the resulting DjVu files to Commons. This is especially so if your PDFs lack an OCR layer. IA will add one and their OCR is quite good these days. Pathore (talk) 21:49, 11 March 2015 (UTC)
That is great information, @Pathore:. To the original poster: If you find IA's upload process daunting, don't let that stop you. Far better than doing nothing, would be to upload them directly to Commons, and let one of us handle the other steps. If you choose that path, hopefully this link is all you need: commons:Special:UploadWizard -Pete (talk) 22:17, 11 March 2015 (UTC)

Should we recreate line diagrams or use original scans?[edit]

I've been working on some 1911 Encyclopedia Britannica articles, and have run into items that straddle the boundary between image and text: diagrams consisting only of lines and words. In one case I decided to recreate the diagrams (see Page:EB1911 - Volume 28.djvu/1059 and below left) which enhances readability, and in another I simply uploaded scans of the diagrams (see Page:EB1911 - Volume 28.djvu/1060 and below right), which preserves more of the original "look" and typography, but at the cost of some readability. Are there best practices for handling such diagrams? Addendum: An ideal situation would be to somehow have all words in the diagram machine-readable, even if displayed at an angle. Thanks. -Animalparty (talk) 22:53, 11 March 2015 (UTC)

If you're redrawing a diagram, it should probably be in SVG. Pathore (talk) 23:42, 11 March 2015 (UTC)
I agree, and have done so. (Note: added galleries above for convenience). Animalparty (talk) 00:31, 12 March 2015 (UTC)
Pictogram voting comment.svg Comment We are reproducing the book, so we should show the figure as it was at the time, with what you are suggesting is an annotated form. We can use alt= within an image to capture words. It is a less than perfect situation either way, and one that harks back to the whole annotation argument or replication vs. modernisation. — billinghurst sDrewth 00:56, 12 March 2015 (UTC)
I had always understood that some degree of modernization is inevitable on Wikisource—after all, we are working in hypertext here. That said, I'm not sure where exactly the line should be drawn, since some diagrams (like these) really are little more than text arranged in an unusual way. We make no attempt to preserve every last exact detail of typography. For example, we certainly don't care about reproducing the typeface of an original in the body text. For the examples Animalparty has provided, how are these simple diagrams significantly different from tables? Both are primarily special arrangements of words, which we often try to duplicate for a table, but we make no effort to ensure that the fonts match the original in a table. An SVG also has the advantage that text inside the figure may be selectable. On the other hand, redrawing even these simple figures could be a large amount of work or require different skills from ordinary proofreading. For figures like these, I propose we consider these figures an example of advanced typography: SVG vector drawings are certainly appreciated, but a simple image taken from the original is also acceptable. Pathore (talk) 01:44, 12 March 2015 (UTC)
I think no formulaic answer will be possible; there is always going to be some judgment call about where the line should be drawn. The nature of what we do with text eliminates the original line breaks, alters footnote styles and sometimes numbering, etc.; so it's hard for me to see merit in the position that our (insert:overriding) goal is to reproduce what was published in a pure form. Given the examples above, assuming that is the best scan likely to be available in the foreseeable future, I would strongly support recreation. That scan looks bad on screen, and would look much worse when printed -- almost illegible. Providing a recreation that could look decent in print offers a valuable asset to the reader, which is not available in the raw page scan, is not available at the Internet Archive, etc. I see this kind of recreation as part of the core value proposition of Wikisource. -Pete (talk) 02:03, 12 March 2015 (UTC)
The sentiments Pete expressed above summarize the situation well. There is no one answer that will always apply. The very fact that we have templates like {{custom rule}} which mimic, but do not exactly reproduce, simple graphics textual divisions clearly illustrates that we are not always simply reproducing the graphical aspect of books. That said, we should try to mimic the source as closely as possible within the limitations of our electronic medium. The easy accessibility of a side-by-side comparison will attend to the rest. --EncycloPetey (talk) 02:17, 12 March 2015 (UTC)

Thanks for weighing in, everyone. I agree there's no firm rule, and it's best to treat these on a case by case basis, and believe that in this case the diagrams are more akin to tables then to images: both tables and these line diagrams represent organized relationships between entities. Had they been a little more "artistic" (e.g. a, unique hand-drawn sketch like File:Darwins first tree.jpg), I would likely have erred on the side of using the original. I think I may upload both original scans and svg versions to aid in verification, in the vein of File:Darwin divergence.jpg and its svg counterpart File:Origin of Species.svg. Lastly, I'm surprised no one added "no pun intended" after discussing "where to draw the line"! Best, Animalparty (talk) 02:38, 13 March 2015 (UTC)

Can not find my book (PediaPress)[edit]

I created a book yesterday named "MEDIVENTURE" in wikipedia but cannot retrieve/edit/find the book anywhere in wiki. The book is still visible in PediaPress but can't be edited (can not insert any new content). I am also not being able to add pages to the book I created. Please help. unsigned comment by Priyambiswas (talk) .

Are you sure it was made/saved here on en[glish].Wikisource.org under the same account? Your User: history & the logs indicate the above was your first time visiting here is why I ask. -- George Orwell III (talk) 09:50, 13 March 2015 (UTC)

Wikisource displays with the wrong resolution[edit]

There is something wrong with the english wikisource site display. It is outsized! This is a problem only here, but not on any other Wiki site, Wikipedia French Wikisource, etc., or elsewhere on the web.- The attached file is from Firefox on Windows 7, but I also tested Linux and it's exactly the same. Would this be the result of the latest wmf update??? Ineuw (talk) 20:51, 13 March 2015 (UTC)

How about now Mr. Modern skin user? -- George Orwell III (talk) 20:56, 13 March 2015 (UTC)
Much better, in fact it's normal. Thank you. What's wrong with the Modern skin? It's offered and I like it. Ineuw (talk) 21:30, 13 March 2015 (UTC)
It's one of the 3 "leftover" skins that was kept with the introduction of the -- and I use the term lightly -- current standard; Vector. And, as you already know, Modern is just a more intricate rehash of the previous standard; MonoBook. Please remember - they both come from a time in Wiki development when the specifications they were based on are nearly all deprectated in today's terms. Plus they existed in a core Wiki environment that wasn't on a weekly schedule for updates like we have now. Bottom line -- just because its available doesn't necessarily mean its fully supported. Even Vector's days are numbered as Visual Editor & Mobile Mode usage looks to overtake desktop view sooner rather than later.

That said - it was totally my fault. I keep forgetting not to base post-change assessments based on how the Special: preference page renderings look. My bad again. -- George Orwell III (talk) 22:08, 13 March 2015 (UTC)

Another problem[edit]

The images on this page are also affected by the size problem. To me, they appear 10 times as large as they should be. Ineuw (talk) 21:42, 13 March 2015 (UTC)

Should be OK now as well. Just another example where one thing works under a given browser & skin but is not necessarily true for all browsers under every skin. -- George Orwell III (talk) 22:10, 13 March 2015 (UTC)

Index:Dictionary of Christian Biography and Literature (1911).djvu[edit]

I have recently more or less finished adding at least basic individual pages to the above index, in the hope, eventually, of transferring the content to the existing Dictionary of Christian Biography and Literature to the End of the Sixth Century. However, I do see a few problems, of which the primary one might be the existing names of the individual article pages which have already been created. Those pages have names often substantially different from the names given to the individual articles in the source itself. Some of those pages may even have names which are problematic in referring to the subject by names by which it is not generally known. What would be the correct way to deal with such matters? John Carter (talk) 22:14, 13 March 2015 (UTC)

It might help if you gave two or three specific examples (with links) of the kinds of problems you anticipate. --EncycloPetey (talk) 01:35, 14 March 2015 (UTC)
Yeah, should have thought of that myself, shouldn't I? So I'm not bright sometimes.;)
Dictionary of Christian Biography and Literature to the End of the Sixth Century/Georgius (43), patron saint of England is a combination of the name of the article as per the bolding, which is "Georgius (43)", with the addition of the first descriptive words of the article. But in that case the subject's common name is generally w:Saint George. Would it be better to use the original name "Georgius (43)", or the existing term, or Saint George?
Dictionary of Christian Biography and Literature to the End of the Sixth Century/Alexander, of Alexandria is evidently, according to the wikipedia article anyway, w:Pope Alexander of Alexandria, although I think the more common name outside of the Coptic church might be Alexander of Alexandria. Which of those names would be best, or, maybe, would Patriarch Alexander of Alexandria be best?
Dictionary of Christian Biography and Literature to the End of the Sixth Century/Joannes (504), abbat of Mt. Sinai is apparently according to wikipedia most generally called w:John Climacus. Would it be best to use "Joannes (504)," which is really kinda weird if you ask me, but the real name of the article, the current descriptive name, which isn't that much more clear, or the name John Climacus?
Those are a few examples. FWIW, my own choice would be to keep the titles of the articles exactly as used in the original source itself. From having looked over the pages of the book, it refers to several other reference books in its bibliographies and several of the articles directly refer to other articles within the work. That being the case, it would probably be easiest to find them if the titles were kept as originally published. And, on a mildly related point, this work says in the introduction it is a collection of articles in a previous edition, with apparently most of the articles omitted. The numbering however is kept consistent with the first edition, and, again according to the introduction, the texts of those articles was kept as consistent with the earlier edition as possible. So, in the unlikely event anyone ever tries to add the earlier edition, it might be easier to find which articles are reproduced verbatim by following the consistent naming patterns of the work themselves. We can always include portal links of some sort to the articles which are perhaps not under what would today be the most obvious names. John Carter (talk) 15:17, 14 March 2015 (UTC)
I would support having all articles under their original names (as found in the scans). We can always have redirects from more "common" names. Pathore (talk) 01:38, 16 March 2015 (UTC)
That leaves the question what to do with the existing pages for the several articles under the somewhat original names. John Carter (talk) 18:26, 16 March 2015 (UTC)
The simple answer would be to move them to the exact names in the source. We can always have redirects. Pathore (talk) 04:26, 17 March 2015 (UTC)

Seeking Advice/tips on Table formatting[edit]

Hi all, I'm trying to figure out how to create tables that only have borders on select columns and cells, case in point: the complex table at Page:Wood 1865 - The Myriapoda of North America.djvu/25 (I'd like to get rid of all the row borders to better match the original table, and if there's a way to automatically center all columns, that would be helpful to in order to avoid manually entering "align=center" on every cell!). I've seen similar tables in Encyclopedia Britannica sources (e.g. File:Zoological Distribution Molluscan regions EB1911.png, so any tips or intervention are greatly appreciated. Thanks in advance, Animalparty (talk) 22:35, 15 March 2015 (UTC)

I decided that making an example by improving that table would be helpful, so I've redone the table formatting on that page. You seem to be looking for the {{ts}} template, which provides shorthand codes for many common bits of table style markup. It also uses CSS, which means that some (but not all—use that preview button!) styles can be inherited. In this case, that means that {{ts|ac}} at the top of the table is enough to make the default alignment for all table cells "center", such that it can still be overridden when needed.
I would suggest reading about the HTML table model, since wikitable markup ultimately expands into XHTML. This can be particularly useful for understanding how to set column widths. (Hint: set a width on the first cell in the column; the rest will (usually) follow suit) The help pages, both here and at Wikipedia describe the details of wikitable syntax. Pathore (talk) 01:35, 16 March 2015 (UTC)
@Pathore: Thanks a lot for making those changes; I'll study the syntax and link for future reference. Cheers, Animalparty (talk) 04:17, 16 March 2015 (UTC)

How would you create a table of contents for The Philosophical Review[edit]

I've had a go at proofreading Volume 12 because I wanted to read an article in it... Any ideas for formatting the table of contents? Cheers, Zoeannl (talk) 02:39, 17 March 2015 (UTC)

The scans strongly resemble the layout used for the List of Illustrations in our current Proofread of the Month. Similar markup to what is used there should work well. Pathore (talk) 04:32, 17 March 2015 (UTC)
That looks neat. Can you tell me where the help page for the markup is? like- float right and dtpl, dotend, djvupageoffset, and djvupage- so I can have an idea what I am doing? Zoeannl (talk) 07:57, 17 March 2015 (UTC)
See {{dotted TOC page listing}} and {{float right}}. Hrishikes (talk) 13:43, 17 March 2015 (UTC)
The preview display has list of "templates used in this preview" below the edit box. Each template link there points to the documentation for that template. Pathore (talk) 21:15, 17 March 2015 (UTC)
I looked and the templates weren't listed? Zoeannl (talk) 07:42, 25 March 2015 (UTC)
Do I rearrange things for the index page so that the pages are listed in order? Zoeannl (talk) 07:57, 17 March 2015 (UTC)
No, preserve the original order. The text should match the scan. Pathore (talk) 21:15, 17 March 2015 (UTC)
Sorry, I meant for the Volume Index page - as done for Popular Science Monthly? Zoeannl (talk) 07:42, 25 March 2015 (UTC)
I don't know what you are asking. If it was transcribed, it should match the scan. If you're referring to making a list of volumes, then those should be in numerical order, I think. If you mean the Table of Contents on the Index: page, then I don't actually know. I've always built the Index: TOC by transcluding text from the relevant Page:s, which would make it match the original order. This seems like less work right now (get both those pages and a TOC for the index at once) and a sorted TOC can always be made later, if needed. Pathore (talk) 05:22, 28 March 2015 (UTC)
The dotted TOC has been working very well, thanks, but I have a problem with articles that have 2 pages listed e.g. Laurie, S. S. 364, 590 Any suggestions? Zoeannl (talk) 07:42, 25 March 2015 (UTC)
Try {{nowrap}}. Moondyne (talk) 08:11, 25 March 2015 (UTC)
@Zoeannl Will have to fix those 2 links after transclusion, they'll need to be hard linked using the "pagetext (3)" parameter.--Rochefoucauld (talk) 12:40, 25 March 2015 (UTC)
Should I leave a marker like {{Missing link}}? Which doesn't exist?? --Zoeannl (talk) 20:38, 25 March 2015 (UTC)
@Zoeannl: For 2 pages listed in a TOC line, the dotted TOC template is quite sufficient. See the list of illustrations at Researches on Irritability of Plants. Hrishikes (talk) 02:50, 28 March 2015 (UTC)

Should complete newspapers be uploaded when only 1 or 2 articles are significant?[edit]

I've been adding some articles from early 20th century scholars concerning evolution and religion (see Evolution and Theological Belief and The Theory of Evolution as an Aid to Faith in God and Belief in the Resurrection) that are historically significant (see e.g. [2]), yet that were originally published in a university student newspaper which contains (at least in my opinion) large amounts of trivial entries like local interest stories, jokes, ads, and general campus gossip (see for instance here) which I have no intention of ever transcribing, although some might conceivably have some future value. I intend to add more articles, and while so far I've been only uploading the relevant pages, I'm wondering if it is worthwhile to upload entire editions of the paper, as I believe I read somewhere that full sources are preferred to excerpts. I'd prefer to not obscure the titles by e.g. retitling them all The White and Blue/Evolution and Theological Belief, and it seems from viewing Category:Newspaper articles that there is no agreement on prefacing articles with the Newspaper name. Any best practices or advice welcome. -Animalparty (talk) 19:47, 20 March 2015 (UTC)

The Index should be the complete issue of the edition. Then just work on the sections of current interest. The other parts will be got to eventually. In terms of titling, redirects are cheap. I suggest using the full sub-paged title (e.g. The White and Blue Vol. xiv No. 12/Evolution and Theological Belief) as the transclusion point and setting up a redirect from the non-prefixed title. I'm pretty sure that there are other articles with Evolution and Theological Belief as the title out there, so when they are added the redirect can become a disambiguation page without the trouble of moving the transcluded articles at that point. Beeswaxcandle (talk) 20:13, 20 March 2015 (UTC)

what happens when index file is changed?[edit]

A hypothetical but realistic question: what are the guidelines or protocol for changes to source djvu or PDFs on Commons? Say for instance a document is missing a page or two (as I've encountered on some Google Books PDFs), and/or contains missing images or incorrectly ordered pages, or if I just want to upload the same file with all "digitized by X" watermarks removed? If such a file is indexed, and then overwritten by a new, better version, what happens? Does a new index need creation? Must the new file include the same number of pages? Similarly, what's the procedure for "Frankenstein" articles, that for instance draw from 2 or more different sources (e.g. a djvu missing a cover image and title page supplemented with a cover .jpg and single page.)? Animalparty (talk) 22:06, 25 March 2015 (UTC)

Rule 1: for a new Index always check the file for missing images or incorrectly ordered pages before starting proofreading.
In case an old Index, if you realize something is wrong, fix the file by inserting/removing the missing pages, newly upload the source file, adjust the pagelist and move the Page:yyy.djvu/n pages as needed. If they are many, ask for a bot request, specifying your needs. See e.g. Wikisource:Bot_requests/Archives/2014#Picturesque_New_Guinea_Bulk_Moves. No new Index pages needed, but transclusions in Main ns and all links affected by the move need to be updated.
If you just want to upload the same file with all "digitized by X" watermarks removed, just upload the new file. But make sure number and position of the pages is unchanged.
Hope it clarifies a bit.--Mpaa (talk) 22:26, 25 March 2015 (UTC)

How can this page be re-created ?[edit]

I was just editing or creating this page and received the response below:

The revision #0 of the page named "Page:History of the Press in Western New York (1847).djvu/35" does not exist. This is usually caused by following an outdated history link to a page that has been deleted. Details can be found in the deletion log."

I don't see anything in the deletion log. Perhaps I clicked on the wrong item. How can this page be re-created ? Thanks

Robin2014 (talk) 16:40, 27 March 2015 (UTC)

I can't see any problems now and can open the page, which I see you were able to save before you posted here. If you're still having problems, I suggest clearing your cache. Doing so often solves this sort of database niggle. Beeswaxcandle (talk) 18:10, 27 March 2015 (UTC)
Excellent ! Thanks. Robin2014 (talk) 20:04, 27 March 2015 (UTC)

Repairs (and moves)[edit]

Other discussions[edit]

File:The International Code of Marketing of Breast-milk Substitutes.pdf[edit]

The license tag on this needs updating... suggested new tag? ShakespeareFan00 (talk) 19:08, 28 February 2015 (UTC)

[Global proposal] m.Wikisource.org: (all) Edit pages[edit]

MediaWiki mobile

Hi, this message is to let you know that, on domains like en.m.wikipedia.org, unregistered users cannot edit. At the Wikimedia Forum, where global configuration changes are normally discussed, a few dozens users propose to restore normal editing permissions on all mobile sites. Please read and comment!

Thanks and sorry for writing in English, Nemo 22:32, 1 March 2015 (UTC)

Tech News: 2015-10[edit]

16:41, 2 March 2015 (UTC)

Implications of changes[edit]

@George Orwell III: Looking at the section link aspect, is there anything that is going to impact (positively or negatively) or in the "lessons learnt" space from this/ I note the discussion at mw:Requests for comment/Clickable section anchors. I also note that this hits us with the imminent release. — billinghurst sDrewth 22:47, 2 March 2015 (UTC)

I was under the weather most of last week so, while I was curiously monitoring this particular development string previously, the brain fart since has left me out of the loop when it comes to the current state of affairs. I will endeavor to "see" what's up on this front over on test2.wikipedia.org which hosts a rudimentary version of our namespaces and the PR extension later tonight - hopefully this change is already live there. I'll touch back afterwards (and nobody is preventing anyone from starting without me either). -- George Orwell III (talk) 22:55, 2 March 2015 (UTC)
@Billinghurst: See https://test2.wikipedia.org/wiki/Headings#top The new "section anchor link" should appear when your mouse hovers to the right of a section title.

I can see the first possible pitfall already - the new anchor "eats" even more of the left-hand gutter/margin where the embedded page-links in dynamic layouts usually reside. If it does, we can start thinking about a collapsing side-bar (or even better -- something lightbox-ish) and build a permanent "gutter" to host everything from sidenotes to pillcrow markers to embedded pagenum links and anything else you can think of. That way, we might actually gain usable screen-space in the end in spite of any possible new elements taking up more screen-space. -- George Orwell III (talk) 08:59, 3 March 2015 (UTC)

We tend not to utilise section headers in our transcriptions, so I think that the actual impact will be minimal for the number of works, though where someone has used sections, it will have the impact as you note. Sounds like we need to hear more about their next skin and make sure that there is a consideration for Proofread Page. — billinghurst sDrewth 12:11, 3 March 2015 (UTC)
That's a growing issue as HTML5 expands the h1 thru h6 scheme. Eventually, these new heading elements (header, footer, aside, nav & the rest along with the existing 6) will be just as commonly used as the current h1 thru h6 scheme. So the current practice of centering a div for chapter headings is going to come back and bite us in that aspect as well as a few others. IOW; we should have been manipulating the h2 thru h6 tags to do stuff like that at some point (eliminating the need of adding an anchor template to such centered-div chapter headings would have been likely with the advent of this latest refinement for example).

And as far as the "next" skin goes -- just switch to Mobile Mode from your desktop and you can get a good idea of where they are looking to go with the "next generation" of skins. -- George Orwell III (talk) 12:28, 3 March 2015 (UTC)

┌───────────────────┘
I’m showing display options frozen as a black link. is this the reason? Slowking4Farmbrough's revenge 16:40, 4 March 2015 (UTC)

Nah. That's just me trying to turn Dynamic Layouts into a gadget. It worked for me -- until you mentioned this & then I tried without logging in -- then I got a static label too. So it's reverted.

Anybody know how to identify the module dependencies being used for running MediaWiki:PageNumbers.js? -- George Orwell III (talk) 18:58, 4 March 2015 (UTC)

ok, thanks, (works again) i don’t know if gadgetification is improvement; i haven’t figured out how to get ocr button back. Slowking4Farmbrough's revenge 19:10, 4 March 2015 (UTC)
I'm sorry to hear that -- and with all due respect -- everybody who invested some editing time to work through a dozen or so steps has resolved their issue by now. Some of those solutions are somewhere above by now & as well as still on my talk page.

And "Gadgetizing" the 3 or 4 remaining "hard" scripts is a necessity for keeping the future possibilities for us open. Plus Common.js does not utilize the ResourceLoader like the gadget extension does (& that is primary reason for porting over). -- George Orwell III (talk) 19:59, 4 March 2015 (UTC)

hi, i see that page links on display option are black again. you of course understand how it is disruptive, if editors are trying to transclude pages into articles. Slowking4Farmbrough's revenge 13:13, 5 March 2015 (UTC)
I think you're the only one this time - re-worked it last night & its working here. Need some more feedback at least (Browser & Operating system). Anyone else? -- George Orwell III (talk) 13:19, 5 March 2015 (UTC)

Inspire Campaign: Improving diversity, improving content[edit]

This March, we’re organizing an Inspire Campaign to encourage and support new ideas for improving gender diversity on Wikimedia projects. Less than 20% of Wikimedia contributors are women, and many important topics are still missing in our content. We invite all Wikimedians to participate. If you have an idea that could help address this problem, please get involved today! The campaign runs until March 31.

All proposals are welcome - research projects, technical solutions, community organizing and outreach initiatives, or something completely new! Funding is available from the Wikimedia Foundation for projects that need financial support. Constructive, positive feedback on ideas is appreciated, and collaboration is encouraged - your skills and experience may help bring someone else’s project to life. Join us at the Inspire Campaign and help this project better represent the world’s knowledge! MediaWiki message delivery (talk) 19:22, 5 March 2015 (UTC)

FWIW: The English Wikisource also has a regular Proofread of the Month. We select a work (or works) specifically written by a woman to be proofread each August. Works by women can be proofread in other months as well, but for August we ensure that the voice of female authors is included in the rotation. --EncycloPetey (talk) 00:08, 9 March 2015 (UTC)
For what it's worth, last time I checked Wikisource was actually quite good for gender balance. Not quite 50:50 but better than most of its sister projects. The only one that was consistently better was Wikiquote. Not to knock the campaign; just an interesting piece of information. - AdamBMorgan (talk) 19:50, 5 March 2015 (UTC)
@AdamBMorgan:, I'm curious, where did you get this info? It seems that the research around this always has a challenge in identifying people's gender, and in determining the criteria for being a "regular user" etc. (My instinct is that you're right -- it seems like there are plenty of women here, compared the the other projects I work on -- but I'm curious about the source.) -Pete (talk) 23:58, 8 March 2015 (UTC)
Example gendergap chart
@Peteforsyth: It was based on a chart on Tool Labs ("gender gap / edits"; which appears to be down at the moment). It wouldn't stand up as rigorous science but it can give an indication. I believe it is based on individual recorded edits and the declared gender of the user (via user preferences). When it is working it looks like the screenshot example to the right (Wikisource is the purple line). - AdamBMorgan (talk) 22:55, 10 March 2015 (UTC)
The problem with trying to determine the actual number(s) for gender make-up starts with the lack of the User: preference setting. Easy way to see if it's set is to use the magic word to check...
  • {{gender:Peteforsyth|he|she|unset}}, unset
  • {{gender:George Orwell III|he|she|unset}}, he
One of the things we could do/should do in order to improve upon this (& some other Preference related issues) is make April's monthly maintenance focus a User: housekeeping-centric one. We can make sure folks visit their settings every so often & setting gender could be highly recommended or something. -- George Orwell III (talk) 00:32, 11 March 2015 (UTC)
Thanks for the specifics, @AdamBMorgan: and @George Orwell III:. I'm reluctant to sign onto the notion of "highly recommending" this setting -- I have no problem with people choosing one or the other, but I think the pros and cons are too extensive, and too varied for different people, and I'm not so comfortable with the notion of making a sweeping recommendation like that, at least without substantial accompanying information to inform a decision. In my case, I simply have an aversion to putting personal information out there in a structured, machine-readable way. I sometimes do it, but my default is to avoid it. I'm sure many of my family members are irritated that I haven't formally acknowledged on Facebook that they are my brother, cousin, etc. But I simply don't see the benefit to me (or to my family members) of giving that information away in structured format.
With gender on Wikipedia, I think it is a given that the vast majority of people will not choose to enter this info. Perhaps a major software-based effort (like, a big dialog box that strongly requests the information) might change that, but I hope we won't go down that road. I think the reality is that genuine gender-oriented research in the Wikimedia world will need to use social techniques for gathering information, rather than computer-assisted techniques. It's a pain, but I don't see me sharing my info as a viable step toward changing that.
OK, that was a lot, but hey…you used me as an example…so I think I deserve to have my say ;) -Pete (talk) 00:50, 11 March 2015 (UTC)
In another place & time -- I'd be of like mind believe it or not. Unfortunately, I think "we" pretty much showed our hand when we went with Pete, George & Adam in our respective User: name selections during account creation. IMHO, User: name recognition is just as "machine" detectable in this day and age as the gender preference setting is believed to be.

With that in mind, I figured there was little, if anything, to be gained by not setting it and maybe others in the same "boat" will reach the same conclusion if any subsequent recommendation is framed in a similar fashion. -- George Orwell III (talk) 21:11, 11 March 2015 (UTC)

The amount of littering on my street, and in my neighborhood, is astonishing to me. I've lived here a couple years, and never seen a place where it's such common practice to toss every piece of trash to the wind without a thought. Still, I don't participate. My tossing the occasional soda can aside would make no significant negative difference; in fact, it would surely be quickly claimed by a recycler. Still, I choose not to participate. Not primarily because it would make a difference to the community, but because it makes a difference to me. -Pete (talk) 15:25, 12 March 2015 (UTC)
That's fine be me. Nevertheless, the original point remains; development planning is woefully handicapped by the lack of accurate data and this gender profile issue is but one of the many existing problems with using preference settings for sampling purposes. I'm only advocating for better sampling thru more proactive User: participation. In this case, I'm basically encouraging periodical reviews of User's own preference settings and is something we should consider formalizing to some degree imho. Those who are not aware of the setting in question might improve the numbers related to this discussion as a result. -- George Orwell III (talk) 21:39, 12 March 2015 (UTC)
Yes, agreed -- and my core response remains as well; show me a plan that will substantially change those numbers -- i.e., there is reason to believe a statistically significant number of people will begin self-identifying to an extent that is actually useful to researchers -- and I will take notice, strongly consider changing my own practice, strongly consider signing onto the kind of recommendation you suggest. But in the absence of a solid theory of how it will be successful, I am skeptical. -Pete (talk) 21:54, 12 March 2015 (UTC)

┌────────────────────────────────────┘

Well, I really don't know how to assure you something like that actually takes place in the end (mostly because we only have control over this project among the hundreds of wikis out there). We do know the latest data in the chart linked above is ~2 years old and a bit obtuse for our needs here so the logical first step in a series of steps to follow would be to better ascertain the current status of that Preference setting for a fixed sampling (i.e. relatively recent and/or active contributors) as a baseline -- This needs someone knowledgeable with py or the API to get done. After that, its all uncharted territory where the previously mentioned formal Monthly Maintenance task is but one approach to test. Whatever we decide to try will then need to be re-sampled to see if we managed to improve our own numbers or not. In short, the entire endeavor does not seem promising at face value but we can't be certain unless we try. The alternative is far less democratic.

The usual way they manage to "induce" Users to [re]visit their Preference settings of late is to add, remove or modify some aspect or behavior tied to a setting or settings and then "break" something (intentionally, unavoidably or otherwise) related to that aspect or behavior. Once normal or expected operation is disrupted for Users and the "fix" remedying it hinges on x, y &/or z being done in one's Preferences, the changes are adopted pretty quick and without debate by the relevant User: group or groups (again, recently active contributors). Users who haven't logged in for months if not years, however, will stay "broken" until they also do the same. -- George Orwell III (talk) 22:55, 12 March 2015 (UTC)

I don't know if this is even slightly germane to anybody's purposes but as evidenced here, out of 144238 current enWS user-ids registered (no idea how many blockees and auto-non-editing-registrations from sister projects that may include?) 366 have declared as females, and 1475 as male. AuFCL (talk) 04:13, 13 March 2015 (UTC)
Thanks for that very interesting link, @AuFCL:. @George Orwell III:, I'm not asking for assurance or a guarantee -- which would clearly be unreasonable -- but for a more complete theory/plan. My belief is that for the foreseeable future, genuine research into the gender breakdown of Wikimedia projects will need to be done using survey-driven data, rather than computer-generated data. I simply don't see how an incremental uptick in self-identification will change that -- or how even a "strong suggestion" could lead to anything more than an incremental uptick. -Pete (talk) 18:07, 13 March 2015 (UTC)
That's the rub. The gender "setting" does nothing except tag a value with he/him, she/her or null. That's all. One can think of it as a poor man's survey -- which makes sense for an all volunteer, non-profit entity -- or one may not. Unless there is a swell in support for doing one thing or the other, I'll be moving on I suppose. -- George Orwell III (talk) 19:22, 13 March 2015 (UTC)
Yeah, I think we've probably talked this one out (and then continued talking for a while, I suppose!) Always find our discussions enlightening, though, thanks for sticking with it. -Pete (talk) 20:29, 13 March 2015 (UTC)

Commons user attempting to delete files used to verify documents at Wikisource[edit]

Please see commons:Commons:Administrators'_noticeboard/User_problems#Disruptive_and_frivolous_deletion_nominations_by_Ellin_Beltz.

Thank you!

-- Cirt (talk) 03:11, 6 March 2015 (UTC)

Please do not panic. Commons keeps these files as they are within scope. I've speedily closed both nominations, see here and here. Regards, AFBorchert (talk) 07:59, 6 March 2015 (UTC)
Please do not assume something about my emotions. And thank you! -- Cirt (talk) 17:03, 6 March 2015 (UTC)
The phrase "please do not panic" was intended with this in mind, i.e. as a friendly hint that nobody will be hurt and the files will survive at Commons. There is no assumption involved. My apologies if this was misunderstood. Regards, AFBorchert (talk) 18:56, 6 March 2015 (UTC)
Ah, my mistake, I love that book! :) -- Cirt (talk) 20:23, 6 March 2015 (UTC)

Explaining Wikisource, why it's valuable, and how to get involved[edit]

I've long been fascinated by how little-known Wikisource is beyond its core contributors, and I'd like to do what I can to help a broader audience understand it (and potentially contribute).

I just made a 12 minute screencast, focusing on grabbing images from old books. I'd like to make more like this -- for various audeinces (e.g., librarians, Wikipedians, etc.) and documenting some of the more technical pieces of Wikisource.

I'd be really interested to know what other Wikisource people think of this, and her any suggestions for improvement, for future topics, or pointers to existing resources that I may have missed. In particular, I've had some discussion with @George Orwell III: and @Hesperian: recently that relate to this; I would love to hear what you guys think.

Hope this is of interest -- and looking forward to any feedback and suggestions. -Pete (talk) 00:03, 9 March 2015 (UTC)

Done well, such videos could be a GoodThingTM. I'd be willing to help in whatever way I can, but not until June when my schedule will free up. There is already a facebook page for Wikisource, but it seldom has new posts, and I don't know who runs it. --EncycloPetey (talk) 00:17, 9 March 2015 (UTC)


Well done, Pete. I enjoyed your presentation. There are also some works from wikisource placed on YouTube as I and others have done. I added music to those I did. However, I would have also presented a side by side photo of the yellow images with the cleaned images as I do in all of my own works with books and images. They are not printed yellow as we all know and the change to the proper contrast, color, &c is important in presentations like yours. Too, I have one blasted problem with some few images and that being how to remove those lines on an image that mess up the image -- they come from the left and right and sometimes are just splotches. How can the images be cleared of them? Does anyone know? I have done several books and lots of images. There are some excellent books worth saving and a photo "is worth a thousand words" I have skipped good books and went to other other books due to the bad images in uploaded books at various areas on Internet. I recall an excellent book on HathiTrust showing beautiful color plates of flowers that I had to skip because I will not use the messed up drawings in color. So that is a problem that needs to be addressed and resolved. Most everything I can handle when it comes to images including colorizing for personal usage on coffee mugs I use. I can take twisted or bent forward, backward, &c. paper -> image -> and corrected those as well. So I would like to know if anyone on any wiki area knows how to remove those lines traveling to and fro ruining images.

AdamBMorgan also has a site where he was (still?) enhancing wiki works. So too with Erasmo Barresi of Italy using YouTube. Links to the front of Wikisource.org work well. So too with Facebook when enhancing and advertizing Wiki areas.

All of these things presented are not new but neither are they promoted.

Your voice comes across as very clear which is also important. Still, you showed all yellow images while speaking of making wiki areas better. I hope that all of this feedback helps all of us who love wiki areas. Kindest regards, —Maury (talk) 08:42, 9 March 2015 (UTC)

Many thanks for the feedback. @William Maury Morris II:, you make a good point about showing a fully enhanced image. I did not intend in this video to show the full range of what is worth doing on Wikisource, I wanted to focus on the process of getting the original, high quality scan -- because accessing that is essential to any restoration work. But, in hindsight, I would have done well to explain more clearly what I was and wasn't covering. Maybe I could address that with some edits to the blog post. (Unfortunately, I find editing the video rather intimidating -- I think it can be done, but it seems complicated to do it well. I may give it a try. For this, I used Screencast-o-matic Pro -- I have tried a number of tools (see w:en:WP:WikiProject Screencast) and that one seems to be the best combination of cheap, easy, and fairly good. But I will probably continue trying others, and welcome suggestions on that too. A format that is easy to edit afterward would be especially useful.)

My friend Mike, who is not really a Wikimedian but has done some volunteer work for Distributed Proofreaders/Project Gutenberg, gave me some good feedback too, which I mostly addressed with the bullet lists in the blog post. I will definitely consider a screencast focused on image restoration, and if I make one -- or find a good one -- I'd be happy to link to it prominently from the blog post, YouTube page, etc. (One consideration is, I'm not really expert at it. I can do some things, but am not sure I'm the best person to do a demo. I'm also not too confident with free software like GIMP, which is what I'd prefer to use if guiding others.)

I've seen @AdamBMorgan:'s blog, and have been impressed -- though I have not seen it lately. I will probably continue to do these on my own regardless, but would of course prefer to collaborate with others if there is interest. -Pete (talk) 21:17, 9 March 2015 (UTC)

My two cents For what it's worth, I always mention Wikisource to librarians—it seems like a slam-dunk to them. —Justin (koavf)TCM 01:43, 10 March 2015 (UTC)
I also mention it to high-school English teachers. Such a teacher once asked me whether there was a Concordance in print for Moby-Dick. In response, I pointed out that Wikisource had an copy that could be searched electronically for free. It made her AP Literature class that much easier to run, and allowed her students to work on their papers from home. --EncycloPetey (talk) 02:15, 10 March 2015 (UTC)
EncycloPetey, when you think or promote books like Moby Dick please remember that Matthew Fontaine Maury is mentioned in it (also 20,000 Leagues Beneath the Sea) and more than once in each book. The point I make here is those books were based on living persons to assist in book sales and fascination. M F Maury will be found on that easily searchable book (important since hard copies is not searchable) - Chapter44, bottom left, Melville writes of M F Maury and "charts" that people knew M F Maury was working on. Melville declares that it so happens that what he (Melville) has written has just been done by M F Maury of the National (Naval) Observatory when it is the opposite -- Melville inserted what Matthew Fontaine had already been working on and had just completed. This was done, perhaps, to promote Melville's book because Whale Charts did not exist as M F Maury created them -- so Melville writing about "Ahab" pouring over whaling charts comes from M F Maury's work on whale charts. It is the footnote of Chapter 44 which I have used for years as my email.

Image Quality[edit]

You asked about an example of an image on my Talk page. It was among the very images you used in your presentations. Those are all yellow but I also noticed, as I have many times when enhancing images, you showed several examples. I refer to
Wilson Price Hunt
"Wilson Price Hunt" on mouse over)

His image is ugly even though it has been worked on. See those lines in that image? That is what I was writing about. Too, it looks like the fellow has been blurred a bit but I don't know. I don't think a man would look as bad as he does regardless of just the lines problem. It looks like an image that was done on a painting with the canvas lines showing. Anyhow, you spoke and have written about "enhanced images" and this is not an enhanced image as I think of the word enhanced.

I admire your work with the "Screencast" on YouTube and elsewhere and hope that you will work up something on true image enhancement. I fully understand you could not, or rather I could not, do everything in one presentation. The old "Windows Movie Maker" also works well as opposed to "Screencast" of which I am not familiar. They are all tools with which to explore and work with. I did 2 works on Cycles on YouTube using Windows Movie Maker, a very old program. I Used images from WikiSource, created a compilation, and added proper music -- it is important that music goes well with subject matter. It promotes that specific book here on WikiSource because I also added text on colored background and showed links to WikiSource.Org

Kindest regards, —Maury (talk) 10:52, 10 March 2015 (UTC)

ON WikiSource to promote WikiSource The Cycle Industry (1921) by W. F. Grew ( —Maury (talk) 11:31, 10 March 2015 (UTC) )

Two Mixed versions:-

https://www.youtube.com/channel/UCgRFe1NfNJB3811QB8R-4Hg

example of moire reduction; look at file history

(Arbitrary outdent) @William Maury Morris II:, thank you for all the detail. Just a quick response for now, more later: the lines you are referring to are called moire. Worth reading up on. I do know some techniques for reducing it; they work well on some images, not so well on others. It won't work well on this example (short version: because the dots are relatively large. Also, though I refer to this as a photo in the presentation, I'm sure it's a painting, which I think explains the ugliness!) Here is an instance where I (mostly) removed the moire, and then darkened and sharpened the image. I'd be happy to do a short screencast to illustrate what I did. I'm out of time this morning -- more later. -Pete (talk) 15:14, 10 March 2015 (UTC)

I've done a bit of work also on 'de-halftoning' images (see commons:File:John_E._Hurst_Building,_Baltimore.jpg for example). The appearance is an artifact of halftone printing methods), but I don't claim to be at all adept at it. The method I am aware of involves using a (usually very slight) Gaussian blur to blend the dots into a more averaged appearance, and then an unsharp mask to restore an 'in focus' appearance. This usually seems to end up with an image that looks slightly blurry, however. The are probably better methods, and some kind of presentation on how to do it would be a very good thing. Revent (talk) 23:33, 10 March 2015 (UTC)
Yes, that's basically the process I use -- with one addition, in cases where the halftone dot is of similar size to the scan resolution. Then, I will first increase the image size, so that the blurring happens over a more forgiving array of pixels. (For instance, if your scan is at 200dpi, and you have an 80dpi halftone screen, you might want to do this. If you have an 80 dpi halftone screen and a 1000dpi scan, it might not be worth it. I do this more by "feel" than by science, I wish I could comment more specifically on when you want to do this!) -Pete (talk) 23:43, 10 March 2015 (UTC)
Scaling up the image is something that I had not considered... I admittedly have not done this a lot, just for a few images where someone asked because they appeared really bad in use. I agree very much, though, that it does seem to largely be a matter of 'feel' as to what parameters to use in the filters... it seems quite dependent on the exact image. I would say a higher scan resolution seems to have a huge impact on the results, so scaling up the image seems like quite good advice. Revent (talk) 23:53, 10 March 2015 (UTC)
The moire (the technical name for the kind of halftone artifact this is) generally results from scanning (or printing) at a resolution that is not very compatible with the resolution of the dots used in the halftone process. As a general rule, the higher the scan (or print) resolution, the less likely there will be moire. But when the numbers are close enough to each other, things get pretty ugly. (Imagine the patterns you'd see if you look at a spinning bicycle wheel through a picket fence…same thing.) If the moire has already been introduced, a good amount of quality has already been lost; you generally can't fix it without blurring detail of the image itself. But, yes, bumping up the resolution before doing that blurring will help some. (Also, fooling with the radius etc. of the blur function is worth trying. I usually use trial and error to narrow it down to the minimum amount of blurring that eliminates the dots.) -Pete (talk) 00:17, 11 March 2015 (UTC)
(nods) My attempts at this type of thing haven't been dealing so much with moire patterns themselves, per se, but simply cases where the actual 'half-toning' of the image itself created something that looked bad at typical resolutions, due to ending up sufficiently 'zoomed in' that the halftone 'optical illusion' failed. An actual true 'moire' pattern itself would indeed be the effect of an interference pattern between the scanning resolution and the 'resolution' of the halftone... Gimp actually has a filter (descreen) that is explicitly intended to remove actual 'moire' patterns, but I have yet to experiment with it. Revent (talk) 00:54, 11 March 2015 (UTC)
To make what I mean a bit more clear, you can have a moire pattern in the actual fullsize image itself (due to scanner artifacts) or a moire pattern visible in resized images (due to a harmonic with the frequency of the halftoning and the resizing ratio). The method I mentioned would work for 'de-halftoning' a fullsize image (which would prevent moire patterns being visible in resized versions) but would probably be unsatisfactory if there were actual moire patterns in the fullsize version (from the scanner). Revent (talk) 01:03, 11 March 2015 (UTC)
You're right -- thanks for that important clarification. Sometimes it's moire we're dealing with, other times it's simply the halftone dot itself -- and there are subtle differences in how to handle each. Very cool to know Gimp has a filter for moire -- I will give that a try. -Pete (talk) 01:05, 11 March 2015 (UTC)
Actually, @Revent:, I think the Tabitha Brown example above illustrates your point nicely. I blurred it enough to get rid of the halftone, but there is a rather subtle moire pattern in the background. I certainly wouldn't want to blur it to the point of eliminating that, as that much blurring would obscure many of the facial features. But in that case, the moire is minor, so it's not a big deal. -Pete (talk) 01:08, 11 March 2015 (UTC)
Indeed, the original has both halftoning and a faint moire (which is still visible in the final version).. you did (IMO) a very good job of removing the halftoning, but that didn't affect the actual 'moire' pattern itself. (I do agree that it's unlikely that an attempt to do so would be worthwhile.) That image is also a good example of the drawback that you have to watch for, particularly around the eyes... the final image is not actually 'less in focus', but has a tendency to appear so when looking at fine details.. you can counteract that to some degree with an unsharp mask, but it seems to improve the 'low-res' appearance while giving visual 'rainbow effects' on edges in the full scale. It's very much a matter of experimentation and practice, as you noted, which is why an actual video or two demonstrating the process would be great. In general, though, it seems like the 'best' answer in all cases is simply a higher resolution scan, as it then gives better results when removing the halftoning. Revent (talk) 01:19, 11 March 2015 (UTC)

Tech News: 2015-11[edit]

15:18, 9 March 2015 (UTC)

"Layout" option in index-transcluded pages disappeared[edit]

Recently two options have been added to sidebar section "Display options", one to show/hide page links, another to show them inside or beside text (the latter doesn't work), while the layout switcher disappeared. Is there a way to get back old page links and layout switcher? I use Firefox & Chromium (Ubuntu 12.04). --Nonexyst (talk) 17:41, 9 March 2015 (UTC)

You're lucky. I just have a lonely "Display Option" section ...--Mpaa (talk) 19:17, 9 March 2015 (UTC)
Well, now (though I use another computer with latest versions of the same browsers) it sets the layout to every Wikisource page. ----Nonexyst (talk) 08:23, 10 March 2015 (UTC)
i see someone was coding with the layout options. they disappeared for me from the 4th to the 10th. it would be nice to put a notice here, so we wouldn’t be guessing and working around. Slowking4Farmbrough's revenge 21:18, 10 March 2015 (UTC)

Untangling the Web[edit]

Fascinating publication, if anyone hasn't come across this yet.

Could be a useful resource for archival and educational purposes, to add here to Wikisource with various subpages, etc.

-- Cirt (talk) 18:47, 10 March 2015 (UTC)

Input for improving integration of Wikidata in watchlists needed[edit]

The following was a message asked to be distributed to communities, and relates to where the information, that we pull from Wikidata, is changed. For example, interwikis, VIAF, etc.

Hey folks :)

Data quality and trust is what we're currently concentrating on in the development around Wikidata. A big part of that is improving the integration of Wikidata in the watchlist of Wikipedia and other sister projects. I just opened a page to collect input on how we can improve it.

Wikidata:Watchlist integration improvement input

Cheers

—Lydia Pintscher (WMDE), wikitech-ambassadors mailing list

For those who have an interest, it would be great if they could have a look. Comment there if you like, or we can comment here and provide a collective feedback. — billinghurst sDrewth 23:38, 11 March 2015 (UTC)

SUL finalization update[edit]

Hi all, please read this page for important information and an update involving SUL finalization, scheduled to take place in one month. Thanks. Keegan (WMF) (talk) 19:45, 13 March 2015 (UTC)

Tech News: 2015-12[edit]

15:14, 16 March 2015 (UTC)

Index:The American Cyclopædia (1879) Volume XII.djvu[edit]

No scans displayed in page view for me? Anyone care to take a look as it's only this volume which is affected..ShakespeareFan00 (talk) 10:47, 17 March 2015 (UTC)

Nothing wrong with it. — billinghurst sDrewth 12:12, 17 March 2015 (UTC)
Working fine for me too.Jpez (talk) 13:13, 17 March 2015 (UTC)
Strange, because I still can't see pagescans. I get the text in the edit box, and if I click image I can see the page image.

It just doesn't seem to want to load pagescans for me when I click Edit. At least it's not a corrupt file. ShakespeareFan00 (talk) 17:40, 17 March 2015 (UTC)

Can you please rather than saying it works for you specify which OS/browser combination? It's not working on XP/Firefox 36.0.1ShakespeareFan00 (talk) 17:47, 17 March 2015 (UTC)
Not working for me on Firefox/Chrome.--Mpaa (talk) 19:27, 17 March 2015 (UTC)
Now working. No clue what happened.--Mpaa (talk) 21:57, 17 March 2015 (UTC)
Sorry I thought you meant the pagelist wasn't showing in the index page for some reason. Scan resolution in edit mode was set to 0 on the index page. It's working fine now. Jpez (talk) 19:40, 17 March 2015 (UTC)
Is it possible to run a db query to find other Index pages with this issue, it seems a straightforward enough situation to resolve? ShakespeareFan00 (talk) 19:44, 17 March 2015 (UTC)

Page:US Patent 755840 (Bose's Microwave Apparatus).djvu/2[edit]

Erm where did the proofread tools option on the edit toolbar go? I wasn't informed that this was to be removed. ShakespeareFan00 (talk) 19:17, 17 March 2015 (UTC)

You must be having local issues. All looks normal to me. — billinghurst sDrewth 11:46, 18 March 2015 (UTC)

Two toolbars simultaneously ?[edit]

When in Page edit mode I see TWO TOOLBARS, one for the header and one for the contents, with no possibility to close the header as before.

Whoever is f**king around with the toolbars, please stop. This is a completely stupid idea. — Ineuw talk 01:49, 18 March 2015 (UTC)

I also am seeing the two toolbars but the weird thing is that it is only on some pages and also the proofread tools are missing when it happens. Jpez (talk) 11:30, 18 March 2015 (UTC)
Giving examples would help. Also telling us what preferences that you have set, in this case for toolbars. Every Tuesday is software update day, and there can always been unintended consequences with changes taking place in one part that can effect us with Proofread Page. — billinghurst sDrewth 11:49, 18 March 2015 (UTC)
"Enable enhanced editing toolbar" set to enabled seems to cause this for me (XP/Firefox 36.0.1.) Turned it off and the problem went away. ShakespeareFan00 (talk) 12:20, 18 March 2015 (UTC)
Is this the problem of having both toolbars enabled? If so the preferences should be changed so only one of them can be enabled at once.. ShakespeareFan00 (talk) 12:51, 18 March 2015 (UTC)
A very quick check suggests the 'Enhanced editing toolbar' needs to be overhauled to be compatible, it also not displaying the Special Characters box when that option is selected (XP/Firefox 36.0.1) ShakespeareFan00 (talk) 12:54, 18 March 2015 (UTC)
Zoom in/out option is also gone for me. Londonjackbooks (talk) 14:15, 18 March 2015 (UTC)
I also unchecked enhanced editing toolbar and at least I can do some editing now (I can use OCR, zoom in out etc. Before these weren't available.) so the problem must be something with the enhanced editing toolbar as ShakespeareFan00 pointed out. I am using firefox with Linux. Jpez (talk) 17:07, 18 March 2015 (UTC)
@Jpez: I must assume that you enabled the legacy toolbar, or were both toolbars checked? — Ineuw talk 03:29, 19 March 2015 (UTC)
Interestingly, this problem occurs in Windows (7) Firefox versions 36 & 37 and Internet Explorer 11, but not in Chromium (Chrome) or Opera. It also works correctly Linux/Xubuntu in all browsers (FF, Chromium and Opera).— Ineuw talk 03:29, 19 March 2015 (UTC)
@Ineuw: I'm sorry I don't even know what the legacy toolbar is. In preferences in the editing tab I unchecked (Enable enhanced editing toolbar) and (Show edit toolbar) is checked, with these settings I have what I assume is a more basic toolbar which at least allows me to do some editing. I tried enabling the enhanced toolbar today to see if maybe it would work and what I noticed is that the first time I load a page to edit in the page namespace everything is fine. If I refresh the page or click on any other page the problem with the two toolbars and no proofread tools persists from then on. I tried logging out and back in and the problem was still there. I tried restarting firefox and the problem was still there. I tried rebooting my computer and the problem was still there. But as as I said there was no problem with the first page I tried to edit this morning. I also tried upgrading firefox from 29 to 32 today but that didn't help. Jpez (talk) 04:20, 19 March 2015 (UTC)
@Jpez: Sorry for not clarifying "Show edit toolbar" is the legacy toolbar. I assume you are using Windows. The problem is with our Mediawiki software update of Tuesday. It was mucked up somehow. The Mediawiki programmers test the software on all browsers, and this problem may have to do with local modifications to our English Wikisource. The problem is that in other browsers and Firefox in Linux work fine, (see my comments above). There is nothing we can do until someone in the know will look at it. — Ineuw talk 04:35, 19 March 2015 (UTC)
No worries @Ineuw: I'm not complaining I'm happy enough with the basic toolbar myself, I'm just trying to help whoever may be able to troubleshoot. I'm using Linux, not windows and still have the problem. Jpez (talk) 04:47, 19 March 2015 (UTC)

T88027 seems to be the root of the problem(s) along with a few subsequent tweakings/changes. Looks like in the attempt to "log" WikiEditor based edits, some(?) one(?) of the module dependencies no longer jibe. Adding the 'schema.Edit' module for those who customize WikiEditor in some fashion seems to rectify the problem however. I added it to the OCR gadget as well but I'm not sure if that will halt the double loading of the WikiEditor toolbar or not. More when I find it. -- George Orwell III (talk) 09:23, 20 March 2015 (UTC)

Hinky djvus at EB1911[edit]

Most of them still seem fine but Volumes 26 and 27 at least have stopped working. There are still pages (like this one that are proofed and accessible via link but our pages can't get to them. — LlywelynII 03:47, 18 March 2015 (UTC)

It would seem that we have some references appearing with <ref name="nn">. Finding it is harder! Looking and looking. — billinghurst sDrewth 11:10, 18 March 2015 (UTC)
The djvu file is jiggered and needs reloading. EB1911_-_Volume_26.djvu ‎(0 × 0 pixels, file size: 95.2 MB, MIME type: image/vnd.djvu)billinghurst sDrewth 11:20, 18 March 2015 (UTC)
On a separate note re the work. You seem to be doing a whole lot of wikilinks that are not apparently native to the work. Generally the practice that we have had in such encyclopaedic works is to wikilink where there is a q.v. indicator. If you utilise so many template transclusions you will also run into the template expansion limit and it will simply stop. Also, for intermediate pages, there is no requirement to use section tags, we would generally only use section tags on partial pages on the first or last when using <pages>. — billinghurst sDrewth 11:23, 18 March 2015 (UTC)

Tech News: 2015-13[edit]

15:09, 23 March 2015 (UTC)

Text justification[edit]

I've noticed when downloading texts in the past in the epub format and reading them in my ereader that the text alignment was not justified. I didn't pay much attention to it until today.

Personally I don't like reading the main text of a text that isn't justified and it puts me off to be honest, also as we try to remain as pure and true to the original text as possible I think this is a major issue. If text is justified in the original (which I believe most texts if not all are) I think it should also be transcluded as such.

I didn't pay much attention to this issue until today when I did the first transcluding of some chapters I've been proofreading where I saw that the text alignment when transcluded isn't justified. You can see the difference here where I tried justifying the text with the justify template I found. From pages 1 to 8 I've used the justify template in their page namespaces and from 9 to 24 is how the text is usually transcluded. When justified the text looks quite nice when transcluded except that the pages don't blend as you can see when each page commences. Is there a way to fix this?

I had a quick search and couldn't see this issue being discussed before. I think this an issue that we seriously need to look into.

There are positives and negatives. You've mentioned one of the positives: a professional and readable look. However, one of the negatives of text justification is that, when there are images present or when the reader is using a mobile device with a narrow screen, the justification forces giant gaps in the text, which looks unprofessional. This is one area where the change in medium from physical book to electronic one means that we really cannot be faithful to the original. An electronic book cannot (and should not) enforce a certain page width and certain line breaks. We do not have the technology to insert hyphenation and other features that allow justified text on the printed page to look the way it does. If we want justification, my feeling is that it should be an option selectable by the end-user, and not something forced onto every reader. --EncycloPetey (talk) 13:20, 25 March 2015 (UTC)
@Jpez: As far as I know, at present there is no rule here which prohibits justified text. I am not going into the pros and cons, because your primary question seems to be about how to do it. This should not be done in the page namespace, that won't transclude properly. But this can be done in the main namespace, by wrapping the transclusion as under:
<div style="text-align: justify;">transclusion</div>

Hrishikes (talk) 04:40, 26 March 2015 (UTC)

Thanks Hrishikes I'll give it a try, though I agree with EncycloPetey that it shouldn't be forced on end-users. I think it would be good if we could toggle it on off somehow. Personally I don't mind some of the extra long spaces that come sometimes from not being able to hyphenate random words automatically, I prefer this to the left aligned text which just doesn't look right for me and bothers me. Jpez (talk) 04:57, 26 March 2015 (UTC)
I'm not sure that Hrishikes' suggestion is a good idea. The result would be to wrap divs with different text-align attributes inside this overall one. There is a way to carry justified text over between pages in the Page: namespace using the alternate forms of {{justify}} covered in the template documentation. However, my preference is to keep the {{justify}} template for particular layout purposes e.g. for a quote at the beginning of chapter like I did on Page:The red and the black (1916).djvu/265. For general mainspace use, justified text is already built into Layouts 2 & 3. It's just on Layout 1 (the default) that a ragged right margin appears. See Help:Layout for more info. Quite how the various layouts translate into epub I'm not sure. Beeswaxcandle (talk) 06:52, 26 March 2015 (UTC)
Even though I've played around with the layout button I never noticed that it justified the text until now. I was hoping it would come out justified for the epub export as well but sadly it didn't. Hrishikes suggestion worked fine with the text I'm working on and hasn't bothered any of the other formatting. I also tried exporting it to epub and it looked very nice on my ereader which is a nook simple touch. Jpez (talk) 19:26, 26 March 2015 (UTC)

Wikipedia Signpost interview[edit]

I figure to be trying to start a regular "column" in the w:Wikipedia Signpost detailing activity at other WMF entities, maybe initially on a one-story a month basis. The first such piece is an interview with, pretty much, anyone interested in taking part, which can be found at w:User:John Carter/Wikisource column. I figure to eventually start one such userspace page per relevant WMF entity which can be reused for future columns on either a regularly-scheduled or as needed basis. This first piece may not run for several weeks yet, which might not be to our disadvantage, as it gives those involved a chance to put together a piece which they could use to best promote wikisource and/or encourage other wikipedia editors to become involved. Any input is more than welcome. John Carter (talk) 17:10, 25 March 2015 (UTC)

Forth Bridge[edit]

Index:Forth Bridge (1890).djvu is nearing completion - some people have already helped out enormously, and further help with proofreading and correctly transcluding it would be brilliant. There are still some tables that need to be converted at c:Category:Forth Bridge (1890) tables, but that and the plates at the back is the only thing keeping it from completion. Ostrichyearning (talk) 10:03, 26 March 2015 (UTC)

Interesting blog post: VIAF and Wikidata[edit]

For those who have an interest in the linking metadata (authority links) that we utilise on author pages (predominantly) and works, there is an interesting blog post where VIAF discusses their decision to migrate their linking from English Wikipedia to Wikidata. The announcement and the resulting comments are worth perusing.

For Wikisources it should have an advantage as the Wikidata landing page displays all existing articles (where linked) on the Wikimedia platforms, and that it includes any links that we have for English and other language Wikisources. This is instead of someone landing at English Wikipedia, and then needing to stumble upon the enWS link.

Pictogram voting comment.svg Comment This change increases the value of our authors/works being added to Wikidata, and emphasises that we need to continue to develop our strategy for data additions. One could say that we have been somewhat reactive in doing tasks at WD to our works/authors, and pretty much leaving the WD people to determine what we have and how. — billinghurst sDrewth 01:38, 29 March 2015 (UTC)